deterministic it is statistical. the same exempt prompt you put in can output something different every simple time you ask it. so that's a new paradigm for software it's a new paradigm for technology that everyone is going to have to get used to. the number one question is trust. this also happens to be our company's number one value. how we engineer trust into the product and help our customers deploy ai in a safe and secure way that respects their data privacy and security? that is number one. the number two question is companies felt underwhelmed by the earlier investments in the last big trend around the blockk chain and the meta- verse. so this time around in the economic environment every ceo what is the business case for ai? brennan, training on the large

language of models can get really expensive. there has to be a business case in mind with its bright driving productivity increase in the sales out bogus driving efficiency there has to be a business case there. and then the third question and it's a really big one is how do we drive change? especially if you're bigger company the more employees you have for. >> like people management question you can buy whatever tech you want but you also have to change people and train people get them ready for. >> you could mail the trust part you got the best data privacy data security with ai. you can have the business case but then how do you read trained re- skill in upscale all of your employees? that's a very tall order went to the '90s how to teach everybody in your company to use internet. it's very difficult it doesn't happen overnight. it cannot be top-down or bottom-up it has to be both but it has to be the company

enabling a secure environment for employees to test and learn and experiment and get really comfortable with these technologies. just with e-mail in the '90s. then it's incumbent on individual line managers and employees i can't do the job without learning e-mail without using the internet. the same question is to be asked by every employee every american today in regards to ai for. >> there's a lot to unpack there. maybe it lip service and safety issues we've had some discussions around that already today. what do developers have to have in mind when it comes to safety what are the top priorities? and then from the sounds of your answer they don't just have to have safety in mind they have to show that they work. they have to prove to customers they have done that work so what is the best way for developers

or this salesforce we don't just have it as a brand. parks what is that mean? first of all like to think of it in terms of the three ages and the four d's tell people remember the three ages what should ai be? it should be helpful, it should be harmless, it should be honest. there's all kinds of saying sametechnology buys guard rail e you could build and i'll talk more about that in terms of the four d's this is what ai these are the big risks of ai we need to think about again as technology builders and providers as well as policymakers. we want to make sure ai -- were being careful about data leakage and data privacy issue. then disinformation, discrimination, we have to think about displacement. some jobs are going to be created some jobs are going to go away. as a society, policymakers,

community how do we come together and help society prepare? in terms of how the sales force thanks about managing the four d's in three ages we start at the technology layer. underlies all of our ai. it's features like citations. do i know where this particular answers coming from? is it point back to the document and reference the official document that it comes back from. includes a feature like data masking. you may not want to send into a large language model. how do wheat mask that dataquest arc we have toxicity filters. within customer data as you can imagine there are certain fields of customer data that we found ed that tend to introduce bias in its ai models.

i could be sons zip code, their gender, or customers have the ability to choose -- ticket marches sensitive data field that gets withheld from what get sent to the large language model. once the output is generated i have another set of checks through toxicity filters our research group has developed. we also take an audit trail every step of the way. what was provided into the model. this a number of features from engineering that's engineered into orenstein public that's a first level. the second is a policy standpoint. it's important to think about how our technologies can be used under put those governors in place into the platform itself. as the underlying but it's the policy we choose to build into the product.

one is calling a customer of the customers chatting with our policy regardless of which customers using that with their customer i have to self identify as ai. the ultimate end customer who is engaging regardless whether it's messaging or e-mail they know they're doing with ai versus a great person. we disallow use of facial recognition there's a lot of bias. as a product policy aspects. finally our office of ethical technologies and ai have been very proactive in sharing our principles for responsible ai we open source these principles we are trying to teach everyone of our customers and implement that when the product and customer terms of service for.

>> remind me of a conversation i had this morning. i was chatting with an executive he was saying the ethical aspects one of the elements of ai policy should be driven centrally ai affects every part of your organization you nelson ed central parts rolling out ai and training people. would you agree with that perspective the ethics are so central they have to come from the top a concert marking people like yourself and with your customers you want the really senior executives having that discussion from the beginning for. >> it ties back to the earlier discussion this ai change management happen top-down? the certain things like ethical principles guardrails providing a secure environment like the einstein trust for your employees. that needs to been able top-down by the ceo.

>> exactly. nor is it wise for some organizations have blocked degenerative ai entirely just like their organizations in the '90s a block the internet told employees they cannot use the internet because of the risks concerns i don't think that's the right call for most companies and industries it hurts your business competitive long-term. better is to offer a safe environment provide training and resources employees have to go through the training and take the initiative to actually start using the technologies and a great job descriptions of how they spend their time for. >> one alaska politics and policy question before we get into workforce and education. we have had a pretty productive discussion in washington over the past year. genuine people want to be better informed about ai. but we are in the zone of good intentions still. not a huge amount of action. unsatisfied with the conversation isn't where it's going?

does it work competence of regulated? >> i've never worked in washington for a public servant i appreciate everybody, all of you are because it is so important is a lot of work that's been going on. as a layperson i have been pleasantly surprised by the level and quality of dialogue i am hearing especially in this country about the opportunities are in some of the risks. one of the areas where i see it as a big opportunity for us this there's a lot of discussion about what not to do what to disallow it. and certainly especially using an updating laws that we have you heard commissioner gomez reference the cbc that's it existing law it's a perfect vehicle for making sure the robocalls from ai are forbidden or can consumer opted in for there's a lot of discussion

already about what we disallow. there's an even bigger need an equal need to talk about what new policies we need to create. about what we should be doing. k-12 education i have young kids in elementary school and i do not know if they are learning the skills they need to be successful in an ai economy. >> i think you do they are not learning them. that's the point. >> recognized very difficult it's state-by-state there's a lot that needs to be happening for there's a dialogue i don't hear enough about to be just as a citizen i want to make sure walk but we should be to acknowledge so we shouldn't be doing. >> putting president for a day the ceo hat combine forces what would you do to inject more ai or a better understanding and skills of ai into the k -- 12

system but we need to own workplaces as adults forget retraining and some upscaling. what do kids today need to have? >> every job is going to be transformed by ai. for us to help educate the workers of tomorrow every single person every single child in america it needs to learn how to use ai to be more effective. they also need to understand the limitations of ai it hallucinates it makes up facts. being able to more than ever understand footnotes and citations and doing your homework and not just taking it as is. social media was a perfect example of people not realizing these algorithms were manipulating our thoughts and that is an awareness every student in america should go through his understanding this is not true just because i see it on the internet this is not truth just because chat gdp

tells me sober i should have a healthy level of skepticism and i should follow these steps to fact check what is being been presented to me before forming an opinion. >> it is almost like ai literacy as an extension of digital literacy not just go and learn how to code there are other softer skills people are going to need. >> not just learning how to code but also learning how to code that's was going to go next next is how can he say i to be more effective? consider banning check gdp for essay writing how do we teach kids when to use it and still be a critical thinker when you have this amazing tool? snow but you don't know for. >> very. >> semi two in the '90s or the 80s when spreadsheets were invented you are not going to not teach kids how to use spreadsheets. but they have to understand how the underlying macros are formulated they have to understand what it's automating for you in the same goes for ai

and understand the downside of ai and when not to trust it and the third piece is coding i'm a big believer every person, every child in america needs learn how to code it is becoming as important as reading and math. >> one minor child on that point. it would not dispute coding is really useful but one of the leaps forward with the generative ai natural language property. more people who do not have the coding skills have ways to ask things to be created through ai. you thought about coding and how does that connect them to work us as adults what we can expect a change in her own jobs over the next few years? >> if you think the natural ai it goes back to understand the underlying mechanism of how it works. and even if every student is not

actually directly coding it's important for them to understand what's under the hood and how they're getting the answers. going back to let's say the excel spreadsheet or graphing calculator analogy. >> thinking about the workforce more specifically, i am not someone who thanks is going to be massive job displacement some jobs are going to go is going to be massive tests displacement elements of her jobs are going to change and in a best case scenario that's going to lead to us doing hired valued work that's not how it works out us on how globalization worked out. and somehow the internet affected journalism. so i am wondering do you have any insights about professions you think will be effective first or where you see people moving the high-value work? >> i've many thoughts on this. first is we don't know what we don't know and i will share an

example. for many years the conventional wisdom evening technology industry was if you are creative you will be safe from ai. but we are seeing now ai is really good at generating images and video. moving everything all at once they is generative ai to produce most of the special effects in that movie. that movie traditionally would have had nearly a thousand people work on special effects they were able to accomplish that was seven people because of ai. we are not always able to get an accurately predict where ai will have impact and what won't we have to be prepared and continuously check in. that being said i see a lot of reasons for optimism i will share one my favorite examples and early customer that we

partner with an eye work very closely with the developer generative ai product is gucci november of 2020 when there is a delegation me and several others from milan we are meeting at the salesforce tower in san francisco. as one of the first music took postcode we were all a little nervous of each meeting in person again. they had an issue which so many companies to drink open which is being able to hire enough staff for their client services a lot of people suddenly wanted to work from home they did not want to be the contact center. we focused our discussion on how do we address the client demand for wanting customer support given the labor crunch, the labor shortage? it was at that meeting our chief scientist salesforce former sanford ai professor talked about the large language models he and his team had been developing this was a year before chat gdp was released.

it was such a big idea and the customer so forward leaning they said let's experiment let's try it out and see if it works so we've been working together over the last two and half years and the amazing thing is just learning from the client service providers how to provide the generative ai and a way that helps address that menial tasks when a customer calls in and has a question especially if you are a newly hired service advisor razzing then scramble to look up the answer and put the client on hold generative ai is able to automate that and provide the answer right there the client service advisor is able to focus on the clients. focus on deepening the relationship what he found a something amazing in addition to reducing the amount of time to address the port questions it has empowered client service advisors to become sellers and

become veranda storytellers because of ai augmenting them at every step for. >> to the need a lot of handholding because luxury brands are controlling and not being experimented did it take a lot of time? >> we work together, spent a lot of learning luxury brand it's important your brand perfect you can't have it a little bit off and so ai had to have guardrails we had to keep testing we had to verify work with most expert client service advisors first so they could spot when their issues and fix the issues once issues were gone we scaled it out to everyone else. but it's amazing this team of customer support representatives have reinvented their careers and now that you're in the best work of their careers are not just limited to customer support they are branded storytellers they are in sales. it reminds me of the movie the matrix.

[laughter] went he decides he wants to learn kung fu is augmented with that skill in right away he knows the skill that is the way i felt sings and is gucci client service advisors work no one taught them how to so no one told them how to be a brand storyteller with ai helping them all the sudden you're able to unlock this incredible potential for quick sitting up they could build it and would take. [laughter] >> no drugs involved are. >> thank you so much we are out of time and fortunately. but it has been a great conversation, thank you. text thank you. [applause] ♪ ♪ ♪ workforce it.

this runs 20 minutes. ♪ ♪ ♪ ♪ ♪ ♪ ♪ next up tracy owens director u.s. public affairs. [applause] >> good afternoon. as you just heard my name is tracy owens i am the public affairs for the u.s. at sage. i am also very proud alumnus. [applause] absolutely. i am grateful for the

opportunity it has provided me. i am even more grateful and proud of the opportunities it is continuing to provide for students. you will hear a little bit more about that later but before i discuss more about morehouse i like to tell you more about sage. for those of you who do not know sage is a 50 tech company founded just over 40 years ago about 10 billion pounds was 11000 colleagues globally. sage is a market leader in accounting, financial, hr and payroll technology for small and medium-size businesses is about nine and a% of the u.s. economy. we enable them to streamline operations make more informed decisions and be more productive bring business owners and their workforces from repetitive administration tasks and enable them to deliver it more more stc and higher value tasks

ultimately increasing the productivity. while we are globally headquartered in that uk, and the u.s. we are headquartered in atlanta. a rapidly growing entrepreneurial community and we have a number of investments there including a new building that quadruples our square footage space in atlanta. and also a lot of investments and skills programs to support entrepreneurs. our company purpose is to knock down barriers so that everyone can thrive. sage is committed to working together with the right part is to build a trusted and close a digital economy. our ai vision is a world where business leaders have real-time trusted information about their business' financial performance to guide their strategy. the scale and of the economy sustainably we need a diverse high-performing human culture. with the rapid advances in ai which will have a profound and ultimately positive effect on

the work there has never been a more important time for companies like sage to embrace community partnerships to ensure progress is inclusive and that it can be scaled. too often tech is rolled out with a one-size-fits-all mentality even though it can have lasting consequences on those not considered during its development. not having diverse backgrounds at the table risks missing on some of the keenest insight available to make technology work at its best. that is one reason why sage embrace the opportunity to partner that includes my sisters from college as well to procreate the software engineering and conversational ai program providing students of color with more opportunities to use in studying emerging technologies. and provide mentoring and internship programs to make sure we capitalize in the town opportunities. i am glad the doctor and some

students are here today. they would not mind if they could stand for second i would love for them to be recognized. [applause] thank you. thank you. we are also joined by a couple of my colleagues from sage who have been instrumental in making sure these programs get off the ground and are big supporters of our efforts. if you would not mind standing? [applause] thank you. the wider opportunity to participate in broadening impact on diversifying the tech industry can be amplified. sage sees this is not only doing what's right and just also is what we should be doing to build strong communities and removing

barriers to success. the workforce in general and tech specifically bring some ofe greatest brainpower available to create the best services as possible. diversifying tech is not enough. we also believe making sure small businesses have an opportunity to thrive it. particularly those who have had historical and systemic barriers placed in front of them. sage and only once to increase traversing the chexsystems of service customers, went to make sure we are providing services to those who need it most of the marketplace our efforts are concentrated on elevating human work by solving a customer's needs by eliminating the repetitive and focusing on building trust bread building trust requires me of the most diverse thinking available and at the table. ai is set to contribute significantly to modernizing re- envisioned accounting roles. making them more appealing to a new generation, talent from all walks of life.

ai can handle routine tasks like transaction processing and arrow spotting improving work life balance by handling entire sets of tasks that thomas will he. ai also supports work/life balance and remote work. fitting right into the demand for a new workforce that is emerging. diversity intact does not just mean for the technical roles its diversity across the entire company. it all works together to build the best products and services for customers. to be cleared technology does not bond to one company it's far bigger than that. as we know talent will likely not just work for us that likely for a lot of people and a lot of companies in this room right now. it is incumbent all of us participate in building the next great workforce it helps lift all communities i make sure the work we do is the best it can possibly be. in closing i would state sage is

working hard how tech provides for a more inclusive environment that helps to build the next great workforce. partnerships like morehouse are just a part of it. but a very important part. more is needed we are certainly committed to doing more toward that effort. collectively we can make a huge difference. and we look forward to working with many of you in this room on the journey. we hope that we can make a real big difference in how the tech workforce and beyond looks into the future. so thank you very much. [applause] ♪ ♪ ♪ place of welcome back with mayor steven benjamin. advisor to the president and director of the white house office of public engagement. [applause]

clicks thank you so much for being here with us. thank you for making the turbo from 1600 pennsylvania two joints at such a great honor to have your projects thank you, thank you so much i do not live at 1600 pennsylvania i just worked there. >> thank you for taking a trip over for work. on the work you are doing on behalf of the biden administration is a focus on engaging communities like ours the tech industry and other constituents to help drive the priorities of the administration. and we appreciate first and foremost how much you have given us the opportunity to participate in that process and help succeed in the important efforts are driving forward. one that is top of mine for all of us in the tech industry and the chips and science act and the implementation taking place there. i wanted to ask for your perspective on that and audits of the bipartisan infrastructure laws a big part of that.

let's start with chips, the action happening there and what is your assessment of what we can do to support the work you're doing? >> first of all thank you. the two most powerful words in the english language are thank you but my wife says yes, ma'am. i say thank you. [laughter] thank you for having me and thank you for the partnership. those who have not had the opportunity to work with us in the past. those of a senior in the civil society who work every single day's subject matter experts are proximate to the world challenges facing the country. we work together to try to make sure wouldn't rising up the presence debuted two years and

whatnot for one reason. we get critical feedback helps us do our job better. the agenda in its entirety has been successful. on job creation. returning critical supply chains around chips. that's an amazing opportunity i am a former mayor is my title suggests. eight comes up here every other month for infrastructure week.

built hundred 36 miles of road. the bridge project on for 2 million jobs across this country we've seen unemployment under stork loads for under 4% for two years running. we see a really smart placed -based economic development. again comes to chips in particular bringing back critical supply helping create high wage jobs including highways jobs it's allowing us a degree of security. access to the middle class unlike any other time it may be these numbers was 1.2 trillion infrastructure even the american

rescue plan. these numbers represented a leverage. it is a significant investment that actually is being complemented much more aggressively by the private sector in partnership with the public sector. in private sector announcements complementing the public sector investments. and if we do this right it will pay benefits for posterity. by helping to improve the economy in the health of our families. >> one other area of this digital infrastructure investment the administration has been very focused on a new have been very focused on its address in the digital divide. access to broadbent which is an important goal given how indispensable broadbent is an all access to life from education, to medical care, to

future opportunities. congress sent a bipartisan basis to be supported a lot of programs that help invest in closing the digital divide. it will close tomorrow because congress has not started it up back up again. it's one of those programs. can you talk a little bit about the president's interest in continuing this work? your view of how the digital divide needs to be viewed what congress should be doing to help foster more and broader access to broadbent for all americans? >> and what sheriff alan davidson has got all caps her. >> he is not. it's probably the first at responder authority incredible work he has done to help build out 21st century infrastructure.

action access to the injured edge. in every field of life is necessary right now. i think it's moving somewhere and $90 billion or what have you. it's really wonderful partnership with internet providers that manifest itself in the acp program. it cannot be overstated how important this is. not taking new applicants as of this week. the program will expire this spring. you think about this scale and how it got to this point. we are talking 23 million people all across this country. give access to free internet or significantly reduces their internet. we have heard story after story

of a real families, real challenges the real issues that not for this program would not have access to the internet. as long as staff is been around only 29 my people instead of arty 23 million people acp and the wonderful partnership we have responsibilities to be a politician and i work for the government. i can't tell you to do something an advocate but for those of you who have a voice it's really important the same bipartisan leadership that you will do on acp that steps up and make sure we will continue to provide these opportunities for families all across this country. the good thing about it when you reach that kind of scale all across the country it is worth advocating for and fighting for hopefully and prayerfully congress steps up and allows us to continue the program and maybe even continue to expand to

progress the president's been out on the road a lot in recent weeks and months talking at the importance of the superstructure investment. it's obviously bringing a lot of great construction projects to a number of places around the country. it's important for the national security for that success. the importance of this investment when the upgrades to systems. the technology investment the iti members are doing. can we help the rest of the country understand how important these investments are? how much they're helping with economic growth and development creating job opportunities. what's the story we should be out there talking about?

>> in the former of the african mayors association. and the mutual admiration society. i would tell you mayors get it they see the significant investment. but it does help to have partnerships with each and every one of you who are true leaders in this field to make sure everyone understands these are not one-time investments. if you build a thoughtful partnership that allows you to strategically invest for posterity. pandemic are economic or social disruptions. i would build for the future. longer term two-tiered from the

public, private and philanthropic center they cannot be overstated. realizing all cities have something in common. the investments have taken unique characteristics i will tell you if you don't arty have a relationship with the mayor and the city you at work or have significant number of people i would encourage you to reach out to her, reach out to him they are ways you can inform each other to make sure these investments not only affect me and you but our children's children and children after that pickwick saucing on to ask you about the stem education and workforce opportunities. weird tracy from sage talk about

we have a great group from morehouse who are visiting with us. tracy did not mention this. it's responsible for producing more black mail graduates with masters and phd's than any other profession the country i did not just make that up. [applause] though the technology industry is focus with help of people like the doctor and the people at morehouse creating stem opportunities for historically underrepresented communities. the buy demonstration is very focused on this as well of the infrastructure investment we are talking about companies the more skilled workers what's the focus there? >> i was hustling in the back like my teenage daughters do i hurt morehouse and spelman that i hear someone as well? i have a freshman at spelman. my heart and my money.

[laughter] i have to shut up the spelman as well. the investment and the talent pipeline the domestic talent pipeline cannot be overstated. it is important to reach deep and wide and building the pipeline for many years to come. i would encourage you if you're already investing in your company's make sure you dramatically widen and deepen the talent pool of the amazing and young people and i'd say you're missing out. you invests in the build up your jute well it is critical and

every just not mail benefit. allows us to the president said to build congress from the bottom of a metal output if we do that working people do so much better than the middle class they continue to grow and certainly those who do well will prosper as well. so is that's our time up? >> thank you for all you do. looking forward to continuing to work with you on behalf of the president and vice president and get the job done god bless you all. [applause]

from the tech policy summit. this is about 35 minutes. ♪ joining us now, it needs to bits, board member and senior vice president of government affairs at qualcomm. [applause] ♪ nate: hey everybody. it's a genuine pleasure for qualcomm to support this event again this year and i really want to thank the iti team for all the hard work it takes to produce another event. it's clear that american

ingenuity drives research and development empowering us to new to the logical futures. policies that promote r&d and strong patent protections will ensure american companies maintain their technology leadership. the united states is set to play a pivotal role in shaping the standards that will guide the future of ai. just as qualcomm played a pivotal role creating wireless technology that fueled the mobile revolution, we are focused on democratizing the benefits of ai technology by an access to it in the palms of people's hands via their connected devices. in 2023, the ai discussions primarily centered around training and demo for large language models. this year the ai conversation will shift to is the proliferation of generative ai applications and it confirmeds making technology both easier to use. as ai becomes more integrated in

our daily lives, addressing more responsible ai becomes increasingly emergent. it is imperative will work collaboratively to address risks then unintended consequences. today we are showcasing just outside of the world's first chipset designed to solve complex tenant is around transparency and authenticity in digital content across smartphones worldwide. in collaboration with our technology partner, we will soon enable devices that can add content credentials to any image output. whether synthetic or authentic. addressing these challenges necessitates strong collaboration between the private sector and public sectors. who better to guide us in these pivotal discussions than administrator alan davidson of the national telecom and information administration, ntia ?before we welcome him, let's take a moment to upload the

administration's forward-thinking investments, notably the funding and the open would you access network initiative through the public wireless supply chain and innovation fund driving connectivity and innovation. it feels it the cycle of innovation, testing and deployment, and foreign companies to leverage their r&d investments and unleash the power of market competition. additionally we commend its work on the national spectrum strategy, which is a blueprint for insuring american leadership in 5g, advanced and 6g technologies. the shift towards open deployment -- the shift towards an open deployment approach for 5g and beyond highlights the pivotal role of initiatives spearheaded by alan davidson in collaboration with teens at the u.s. department of commerce and across the u.s. government. now with great excitement and

anticipation, let's welcome administrator davis into the state along with the iti president and ceo to share their insights on navigating ai and artificial intelligence policy. [applause] ♪ >> thank you. appreciate the support of qualcomm. hello again everybody. please. >> thank you. >> alan davidson joins us as he did last year and in prior events and we are thrilled to have you back. welcome. >> i am excited to be here, thanks. it's a great space. >> it is a cool space and it's a beautiful day inside and out. thank you for taking the time to talk to us. a lot of new. that is in relevance to the iti membership, nothing on you plate is not in relevance to the membership. >> [laughter] here working on a lot of things.

>> let's start with ai. you are referenced in the executive order and work is underway on the very tight timetable. i think the first 90 day deadline expired on a sunday. can you talk to us about what you have the team prioritizing and what you need from us in order to accomplish that lofty goals for ntia activity? >> festival thank you for having me. as a former iti board member in a past life, it's always great to be back. congratulations on what you have been doing. it's an exciting and interesting time to be working in this space. everybody has been hearing about it a lot already today, that responsible ai innovation will transform our economy, it's hard to imagine a sector of the economy that will not be touched by all of this, and it will bring a lot of benefits to people, that is, i think, the strong belief that ntia has in

the administration. we will only succeed in this if we ensure we are also addressing the serious risks out there, risks that exist today, not speculative risks, but risks around security, privacy, bias in systems. we are thinking across-the-board about all of these things. you have seen with a sense of urgency with the government is addressing this. we are early still in the lifecycle of all of this. people are jumping in, and i think a perfectly. we saw the set of commitments we were able to get from leading ai companies this number, incredibly important. the executive order which people talk about which we are now implementing against, or important, very ambitious. the work we are doing internationally with the g7. nta has been part of all of that. we serve as the president's advisor on technical medications and policy, we are deeply engaged in the ai policy case. for us there are a couple of big projects, we are doing a big

report on ai accountability and auditing which we started over a year ago. we have a big homework assignment on ai models. also, a lot of work at the policy level internationally. so there is a lot going on. we are excited to be a part of what is happening in the administration. >> that is great. i will give the opportunity to talk about open models if you dress a rumor on the record between you and me, that ntia is considering changing its name to ntai. [laughter] >> who is talking about ai these days? why would we do such a thing? he is completely joking. [laughter] i think that would require action by congress to change our name, but a guy can dream. [laughter] we are doing a lot in this space on this issue around open model weights, widely available model

weights. what people would call open-source ai. it's a really hard problem that we have a homework assignment in the executive order to produce a report for the president by the end of july, so we are on the clock, about what our approach to these openai systems should be. there are real risks and benefits in the idea. there is concern about safety. if you open up models and let them be used without the kinds of safeguards that might be billeted otherwise, what are the risks? that the same there can be benefits to openness. there is a concern within the administration, a competition concern that there are only a small number of companies that control the most important models. openness can be an approach to that. this will be an ongoing conversation. the good news is we have learned already that there are gradients of openness.

it is not a binary thing and that may provide us a path think about how we address this. stay tuned for the report coming out in july. >> speaking of things that require further activity and action on report, sector strategies. ntia, as you mentioned, is advisor to the president. occasionally the president has to issue public reminders of that, as you did with the faa, but we will not talk about that. [laughter] what we are going to talk about is the national spectrum strategy. that was a comprehensive, almost overwhelming effort to bring to bring some direction to the future course of activity in this area. you have got -- i don't know how you have enough people to do other work you have been assigned to do, but this one, tell us where the latest area of activity is on that. >> this is a super interesting and important area that doesn't always get the airtime, no pun intended, sorry, there aren't a

lot of spectrum jokes out there. [laughter] --that it deserves. it is probably one of our nation 's most important resources and also one of our most scarce resources. the demand for spectrum continues to increase. we needed for all the connected devices we know and love. the federal government needs it for everything from aviation radars, satellites to monitor the climate, all of that powered by spectrum. we need to do a good job of managing these scarce resources we have. the last two years, we have been working on a national spectrum strategy which we have been overdoing to have. so pleased we were able to release one at the white house this november. thank you for joining us. >> thank you for the invitation, it was great to be there. >> and it lays out a path for how we are going, as a country,

to tackle is very hard challenge about making sure we can feed the pipeline of spectrum needed for private-sector use and also make sure we have enough for our federal missions. we laid out a four-part strategy which includes his line of spectrum to study in the next couple of years, work on new spectrum technologies like spectrum sharing, new collaborations with the private sector, and a real attention workforce in this space. who need more engineers. other kids to want to study machine learning, but there good jobs in understanding spectrum. we need more people in this space. >> so a lot of work to be done on the spectrum strategy. of course this is unfortunately outside of your purview, but congress has not moved forward on a couple of issues relevant to ntia's work. one is from auctions, which many have expressed concern about, auction authority expiring.

what is the impact of that and what are your hopes, recognizing you don't have a voice in congress, what are your hopes for getting that issue resolved? >> it's kind of crazy that we don't have spectrum auction authority for the fcc. >> the chairwoman has spoken pretty eloquently about how important it is for their continued work in the space and it we will continue to have a smart spectrum strategy for this country, we need to have that auction authority reinstated. we are hopeful that there will be an opportunity to do that. full-on both sides of the aisle understand the importance of that. >> yes, let's talk about broadband more generally, another area where somehow you have people on staff who are working on this issue. >> ntia is like a start up government. i have been the chair for about two years, almost 40% of ntia is

brand-new since i started. so we have had to grow a lot for this. a lot of it for our broadband work. but it's fun to run a start up do it in government. we have grown a lot. >> next talk about government. what is the most important, from your perspective, of the work underway? >> it is a historic moment. . we have this incredible opportunity -- we have been talking about the digital divide for 25 years. in the bipartisan infrastructure law we finally got resources to do something about it. the president has given us a mission, connect everyone. everyone in america. he says it all the time. with reliable, internet service. we are well on the way. it will be a years-long project. it will take a ton of work. we have programs in flight, middle mile, $1 billion out the door. to billion dollars in tribal connectivity finding out the

door. now with the $2 billion state grant program is in flight. this year will be a big year. last year we got plans in at the end of the year from every state about how they will spend the money we will be giving them. this year they will implement those plans. all eyes turn to the states, they will start doing their grantmaking. we need everybody's attention. this is an all hands all day, on deck moment. the truth is money is being spent right now to connect people. our broadband infrastructure program, we have passed over 36,000 homes past with the infrastructure we have built and funded through that. there are real people being connected right now. there is a lot more to come. we will not get tens of billions of dollars from congress to do this again so we have to get it right. >> last question. you have touched on a lot of areas of activity relevant to the tech industry,, them are

areas you have been working on for decades at ntia and some are areas you have been working on this year for the first time. for industry, what do you need help with? where are the areas, the gaps in knowledge education, partnership programs, inputs into all of these different areas or areas you haven't talked about, that industry can be helpful and be of service to you as you endeavor to accomplish these goals? >> such a great invitation. i will [laughter] [laughter] say the list is long. you have heard a lot about ai. the truth is it is worthy of our attention right now and i think getting smarter, i just talked about the open model weights project, it's a big city. these are cases of first impression. we need to be smart with the industry. we have a lot going on with the old internet policy issues -- privacy. i am cochairing a task force on kids' safety right now, figuring

out how we can come back with industry best practices. we are doing a workshop in stanford we just announced on march 13, to get industry input on this. that is another area. broadband, we need funding. that is another thing that is important. you heard about that earlier today, i know, just making sure it is all hands on deck effort paying attention. we just need more,, i would say, people who have dual competency in government, who can understand technology and talk about the policy implications. we really appreciate the partnership with folks in industries, folks in the nonprofit world who help us understand these issues better. thank you for your leadership in this space, thanks to ati and everybody here for being part of the conversation. >> thank you. talking about the intersection between the tactical know-how and policy chops you have described yourself. we are very happy to have you

here. please join me in thanking alan davidson. > thank you. [applause] ♪ >> please welcome anne neuberger, deputy assistant to the president and w national security advisor, cyber and emerging technology at the national security council, and barbara hampton, president and ceo of siemens usa. [applause] >> [laughter] anne, so glad we are together. and actually i am so glad that jason has given the siemens, the chance to be your interviewer

for this last segment, i love our conversations. for those who don't know, siemens is a company that has been working in the field of infrastructure our whole history. in the last couple of decades have been about creating the internet of people, the way we entertain ourselves, the way we engage as a consumer, now the world is recognizing that we are introducing the internet of things and it is an order of magnitude larger than the internet of people. so the key questions of cybersecurity and the use of ai are coming to infrastructure everywhere, and here you are smack dab in the middle of all of that. i know we have questions about the executive order, et cetera, but you are in a room where it would be tempting to focus solely on risk. and yet, you also focus on opportunity. tell me about the balance that you are working to achieve. >> absolutely. first, it's great to be here without a view, goods to be here

with you, barbara. one of the great things about iti -- thank you to our hosts -- is it connects companies and government as an act policy, because of the country, to your question, we are the leading innovator on the global stage. america's economy is the largest in the world because we encourage innovation and we really want to be a leader in not only in technology, but technology envelops many fields. barbara and i have had many conversations about cybersecurity, about telecom policy as part of that. when i think about, to your question, the promise and peril of ai, i know it's the end of the day, you have had many discussions here so i thought i would answer it in the context of a few examples of where we see both, to get us to think about how we bring that together. in the area of promise and peril, to your first question on cybersecurity, two areas. these same models that can help

us really analyze, help a coder analyze code to find vulnerabilities, to find holes better and that a malicious attacker would look for to exploit, to improve the code, to make it stronger, to make it harder to exploit in the hands of the malicious attacker, those same models help them find vulnerabilities that have not been patched that have, not been secured, to use those to build more focused malware. that is an example of that. we want to ensure defenders are using those models first. . that we are improving products as we build them before we deploy them, as well as then using those to find full abilities and the most important products deployed across our systems and power systems to patch them first. one example. i think a second one which always touches my heart is, my husband and i have been involved in a charity that deals with individuals with als. people with als essentially lose

their ability to communicate, lose their ability to move. and their minds are fully intact. their body essentially slowly disintegrates. there are various efforts to essentially take their voices and save them so that as the individual is only communicating by blinking or banging their head, that is translated towards so they can actually still communicate we. . know the downsides of voice coding, we saw it in the. ,. at the head of fraudsters, it's a clearer concerns. so we see the promise and peril in a number of areas of ai. our goal is to ensure on the promised side we are moving at speedwell thank you from the onset, how do we do this responsibly to manage the risks? >> fascinating. we are living in a world where there promise and peril go

hand-in-hand. let's pull the thread a little bit more about, in the world of infrastructure, in the past, people used to say, thank heavens it's not connected. but we recognize there are threats to infrastructure, connected or not and, in fact, it might be that connecting and actually using the tools available to us today could strengthen our infrastructure. your thoughts. >> i could not agree more. about six weeks ago, we dealt with the iranian government hacking connected water systems. water systems connected to the internet. they were able to have an impact on those water systems because they used the default password of 1111. it barely deserves to be called a cyber attack. they affected water systems across 16 states. so, to your point, those systems were not likely designed to be connected to the internet, but over time, it was a way to

download information, perhaps a way for an administrator to also get his or her email as they do their job. so, security has to look at the way products are used today and ensure we can secure them appropriately, and then, i have a question for you -- as we build new, have to recognize that autonomous devices will be connected to the internet. think the infrastructure that we have. our electricity infrastructure has components which ideally could begin their health --beaco n their health. for example after a major flood, people can see which lines are down, which lines should be optimized for maintenance. in an ideal world, we want the infrastructure to be connected in a way that is secure, predictable, and in the way that we can make sure it is protected from malicious attackers. so in line with that, siemens bills out so much connected infrastructure, connected factories more broadly.

how do you think about these new risks as new products are built, as well as you think about the broader infrastructure that you maintain? >> actually, thank you for asking that because this is one of the most creative spaces anyone can work in today. yes, siemens is an expert in all kinds of infrastructure, as you mentioned, the power grid, transportation networks, factories, buildings. there is control technology and all those things. and, frankly, people couldn't even envision connected infrastructure, call at four decades ago. so, the idea of the way things were designed previously was fundamentally different. here we live today in a world where we are able to put these tools to use. we have a research and development department that is constantly scanning the environment. we know that the value we can bring today is burning the real and digital worlds together across all of this built

infrastructure. since 1970, researchers have been working with technologies like artificial intelligence, fantastic. the key question is, how do we apply all of these new technologies to the world we live in, we have been speaking in terms of industrial ai. instead of using general ai on all the internet, we can creates trusted data sets using, say, utilities or a manufacturer's data. imagine a world where you have sensors on all your equipment. you are producing so much information, you simply cannot process it all. now with ai, we can. so with about 2000 ai experts across the global organization, with about 1400 cyber efforts working hand-in-hand with developers, we are constantly evolving our design processes and standards so we are actually building these things in.

the concept is technology with purpose, making sure we are using our best imagination to imagine both the upside we are trying to achieve, as well as the potential unintended downsides and how to best build for protection. >> it's really interesting because, when we think about the most critical cyber security risks of the country, what we are most concerned about are malicious countries or criminals hacking or disrupting the water, power or pipeline systems that americans rely on everyday. as individuals and as companies. those power companies, water companies are largely owned by private sector entities. over the last -- doing the biden administration, who fundamentally made changes to the model of how those systems are secured while still working to keep the close partnership between the nation's critical infrastructure companies and the

government. we have a couple of examples of that that you and i were talking about in the green room, that i want to lift up here and get your thoughts on. we are one of the last countries in the world to put in place minimum cybersecurity standards for pipelines, for water systems , to prevent the kind of thing i described where if our government could hack into a water system merely by identifying and trying default last words. then i will note that what was interesting was, there were 16 systems if i recall correctly, that were hacked. all had default passwords. additional systems that didn't were not the actually hacked. they don't cost much money, those basic practices. those are now mandated for some key sectors under existing safety rules. the second piece was involuntary program called the cyber -- program which aims to say,

americans are bringing big monitors to their homes, mart security systems, mart doorbell systems into their homes, fitness trackers, routers in our offices and across other schools, and we need to ensure they are built secure from the heart. so think about energy star for cyber. it's a program where products can have the cyber trustmark and then consumers can look for that and say, we can be safe online. we hope to see products on shelves by the end of the year. so as you think about it as a company, the two models i just described, one is a revelatory model to lift up all boats to a minimum. the second one is to incentivize products --, frankly, it's a wild west that there of the internet of things devices -- to build things more securely and to bring customers into the mix to show that they value secure products as they shop. i would welcome your thoughts on how you think about this.

>> both/and. a risk-based approach is how we think about these things. how are these devices going to be used? we will need varying degrees of assurance. first we set a framework for responsibility ourselves. in cyber, we established a charter of trust. this is what responsible companies should be doing. but we also recognize that we are going to need public trust as well. with that, a risk-based approach with increasing degrees depending on the criticality of the devices we are talking about. we believe that is going to be an essential part of our future. anne, i am watching the time with dried because i can talk to you all afternoon about these topics, and i know that our audience is very interested. two words about the executive order, and let me ask you to sort of wrap in then how do you want, as a government leader, how do you want the private sector to be engaging as an

administration is launching this executive order and and setting a framework for us? >> so, the executive order is step two of part of a three-step process. the first was, the president negotiated voluntary commitments among a key set of ai companies to say, you are building the technology, you have a responsibility to do so in our that gains americans' trust. things like transparency of the data that models are trained on. vigorous red-teaming and standards for that, that we need to be building. as well as working to gether to ensure that when models are deployed, is important to make sure how humans interact with it and how they are further refined to make sure they are safe overtime. that was then followed by the president's executive order, really for folks who see it, a lengthy, vigorous executive order that aimed to hit at both

the items you said, both encourage ai in areas where we see tremendous promise -- government, operations, how to make government operations more efficient answer questions about complex areas in a simpler way. like in the education field, where we want to encourage using ai to help tailor the way kids study because each child studies in a different way. and also, managing the risks. for example, giving guidelines for how transparent ai models need to be if they are being used for job selection or in loans. because there are concerns. garbage in, garbage out, to use an old cap term that i imagine -- old tech term that i imagine everyone in this audience is somebody with. if models are trained unbiased data, they will spit out no surprise, further recommendations likely to reflect that. so those are societal issues. as well as on the broader side concerns regarding how we test , the models for cyber security risks. for how they could potentially

advance adversaries in developing different kinds of weapons that we are concerned about. so the executive order is rigorous and set on a very tight timeframe. what i am particular focused on as it relates to our conversation is how models are deployed across critical infrastructure. so for example, you could see the promise in models used to deficiency to rail systems. how should signaling systems provide guidance. where are the humans in the loop ? where is the additional risk in those that the department of homeland security is receiving risk assessments from the various regulators. we use that as a final step as input to step three which is congress's work to build new laws to deal with the particular risk that arrived from ai. the second work we can do within current law. the third, where new laws are needed. i think as we close the call, it is really a call to action, to

say that as we bring the promise of ai in so many areas, that there is a partnership between the private sector and government, something iti really focuses on. so we are working together. for example, then you nist ai safety center that the department of commerce is standing up, a new leader was announced, is working to define standards for red teaming. we need companies sharing what they learn as they red team so that is put in place to make that testing is rigorous as it is to be. given the important ways in which we will see ai deployed across our economy and across our society. >> we are in the midst of change. can i just sit at siemens, we are so proud of the members of the administration who have put these first two steps in and we are working closely supporting the efforts of leader schumer as he is drawing in the private sector to help shape that future legislation that will be that third step.

these are remarkable times. thankfully, we have remarkable leaders like you, anne neuberger. thank you so much. >> thank you, i appreciate all the work we do together. thank you all for listening and thank you for your time. [applause] ♪ >> joining us now, ashley, executive vice president for public affairs at iti. [applause] >> that was a terrific conversation that we just heard between anne neuberger and barbara compton, which emphasized the relationship between federal government and industry on the future of so many clinical tech policy issues. in what is sure to be a dynamic year cap policy, iti will continue to convene the

necessary and timely conversations on the topics that are top of mind for all of you. we will continue to bring together voices from across the public and private sectors, explore pressing challenges and help to advance opportunities for u.s. tech leadership. i want to say a big thanks to all of our amazing speakers who have joined us today to share their insights about the future of innovation. it has been an honor to have you on the intersect stage. also a huge thanks to our sponsors. these iti-member companies made today possible. they are amazon, salesforce, cisco, net app, cognizant, accenture, batter, sage, siemens, and mastercard. we encourage you to share your thoughts about today's conversation on social media using the hashtag hashtagthein

tersect2024. also on linkedin or on itic.org. thank you to the audience for stick with us this afternoon. please take a moment to complete our post-event survey which is accessible by qr code. it is brief, and will help us optimize our events programs going forward. finally, we have a reception, so the day is not over quite yet. we hope you will join us just outside the back doors for a reception celebrating cyber security innovation hosted by mastercard and net app. there you can also participate in activations by global innovators and network and connect with other tech policy leaders. please join us outside this room immediately following these remarks. again, thanks for being part of the intersect 2024. we will see you. [applause] ♪

migration the economy. this is about 40 minutes.

1.6 trillion dollars, interest outlays are roughly 1.5 times larger than either defense or nondefense spending. also boosted deficits are two underlying trends, the aging of the population and growth in federal health spending per beneficiary. those trends put upward pressure on mandatory spending. measured in relation to economic output, federal debt by help --

held by the public rises to 99% in 2024 to 160% in 2034. surpassing its historical peak. and then the debt ratio continues to rise, reaching 172% by 2054. from 2024 to 2033, the deficit is about 7% smaller than we projected last year. primarily as a result of the fiscal responsibly act of 2023 and the subsequent continuing resolution. together, those laws produce the discretionary spending. in our projections, the deficit is also smaller than it was last year because economic output is greater. partly as a result of more people working. the labor force in 2020 -- 20 23

is larger by 5.2 million people. also because of higher debt immigration. that in turn leads to additional tax revenue. as a result of those changes in the labor force, we estimate that 2033 to 22 -- 2033 to 2034, gdp will be greater and revenues will be greater by about $1 trillion that they would be otherwise. we are continuing to assess the implications of immigration and revenue spending. two key factors partially offset that deficit reduction relative to last year's projections. the interest cost rises as part of the interest rate and cost of energy related cash provisions are much higher than the staff of joint committee originally

projected. this includes market developments and actions taken by the administration to implement the tax provisions. turning to our economic projections, the u.s. economy grew faster in 2023 than 2022. even as inflation slowed. economic growth is projected to slow in 2024 amidst increased unemployment and lower inflation. starting in the middle of the calendar year. in our projections, economic growth rebounded to 2025 and then moderated in later years. since february of 2023, the agency has lowered the projections of economic growth and inflation by 2024.

they expect interest rates to be higher than projected last year. i am happy to take questions. i would be grateful if you could just say your name and his organization when he asked the question. yes quick the deficits would be down about 1.4 trillion in the because of fra?

>> it is offset by a range of other things, the tentacles such as for the tax provision for the 2020 reconciliation legislation. that leads to a higher deficit or range of changes in each direction. a full discussion in chapter three of the report starts on page 76. it goes through all the gory details. quotes i noticed 2017 was mentioned about 20 times in the report. there are references that come down to tcg expiring. quest that is correct. you can see those effects and revenue figures.

from 25 to .26, the revenue share of gdp goes up. that is attributable to the personal site provisions in the 2017 act. >> good to see you again, i am with reuters news service. just trying to get through the clean energy tax credit adjustment here. you have a $128 billion figure which includes i think the epa new estimates of gas tax revenue from epa changes. could you break out what the inflation reduction act part of this is? how much are those estimates higher than what gcp came out with? >> i am flipping pages. you'll see on page 86, it is

almost two full pages that goes through the details. the challenge here is that it is a new baseline. as compared to the baseline of which the 2022 act was estimated against. originally with jct. two baselines go because the baseline was estimated with the 22. that is a whole other change. makes it a bit difficult to have an apples to apples comparison. but i can discuss some of the key components in the biggest in the want is epa rule. an ep able to has so far been proposed that would take effect with the 2027 model year. in our baseline, that is in its

with have strengths. that is the normal procedure for a rule that has been proposed but not yet finalized. that would lead to -- lead to two main things for the budget. one main driving force that would shift both producers and consumers more heavily toward electric vehicles and that would affect the cost of the tax credits for electric vehicles and that would affect the excess tax for fuel. that is the biggest single piece. >> there was 150 something billion? what was the figure? >> there are lots of things going on on the tv side. we will see if john mcclellan

might jump in as well. on the ev side, the other thing that happened is the treasure implementation of the ev tax credit was different than what they expected them what was embodied in the original cost estimate. then that has to do with some of the basic provisions. the limitations for the tax credit and the legislation for an individual by vehicle. those limitations did not apply. greater use of the ev tax credit. it is just in market development. the the industry has expanded by more than the original tax credit.

i know you would ask about some of the numbers. would that be helpful as well? quest you want to go through some of the things i hear to the extent that you can? >> aggregate costs here -- about two thirds of that are on the revenue side. dirt revenue includes the taxes but also the digital corporate taxes. we consider all of the factors at once.

all the makers say they will promote adoption of ev's. we don't separately identify the portion of ev's. as a whole, it is $224 billion of the budget. there was the issuance of this regulation. quick the other side. that is a separate piece. that gets you to the 428 billion. >> correct. quick thinking about how many ev's would be sold each year.

that is what has to do with the fuel efficiency. but it is difficult for us to disentangle how much that change is due to the epa. especially after all the things that have changed in the economy . >> yes. some of this is on the revenue side. that is an additional complication. chris to follow-up on the, i know you can't really compare. the order of magnitude is -- clean energy credits are roughly double. >> what a way to think about

this. you would anticipate that the tax expenditures are point large now. questionnaires kind of a halfway prevention of the proposed rule. there will be another adjustment that increases the deficit. if the next administration were to come in and undo. >> that would show up. hopefully everyone got that. if you didn't, we can go through it again but but that could show up in different ways. the legislation to undo that rule. it is only halfway in the baseline.

only half the costs are in the baseline so amazing it would be saving people half the cost. if it is finalized, then there is legislation to repeal it, that would have full savings. because there are other regulations where you're using the commission. >> i can't think of others. >> they themselves don't have that big of an impact. dpa is somewhat different than what was described.

>> can talk about the that interest costs? i am not sure i point understand. it doesn't look like we are projecting that in the long run for the industry to be different. i guess what i'm saying is the term premium had gone up. to say that that episode will conclude that in the future so that premiums will be more elevated -- does that make sense? >> i got it. everyone else got it.

if anything is unclear, we can come back to it. we do have higher rates in the forecast embodied in these projections than we did in the previous budget projections. of the impact on big interest outlays. you can think of that as two thirds resulting from the higher rates. there is a lengthy footnote in the report that goes from the two thirds, the one third. both the we can find it. if you look at our economic chapter, chapter two, hopefully that is very handy for everyone by the numbers document. you can see we do not get paid by the click for this. i hope everyone including our viewers on c-span be able to

download it and look at it. it is a very handy document. you can see we do have higher rates. the profile is at the beginning of year. we had the tenure at 42 and then it rises over the course of the year. as of early today, it was around 41. the beginning of our budget update is not exactly perfect but it seems pretty good for the interest right now. this is over the near term horizon. even as the fed projections have paused interest rate hikes, the short end of the curve starts to

lower the fed funds rate around the middle of the year. the first week in our forecast was in may. even short-term rates are coming down and we have long-term rates coming up. that is modestly over .4 and 25. it is partly the term premium and then partly more debt driving rates up and then inflation continues to moderate back down to the fed target. then we see the long-term rate coming down. to go back to the question you asked, you are right, our long-term rate is still pretty moderate. >> we haven't really changed that in the of what happened.

>> that is a fair way of putting it. >> i was just intrigued by the immigration and how much was contributed to your projection. to what degree is this pent-up supply if you like? that will continue into the future years, the projections. also, you touched on the immigration. you will be researching that. what would be a well-designed immigration? >> let me take those in turn. in chapter two of the report, there is a long box on the impact of immigration on the economy and then from the budget. as claire said, it is a salient

impact in our economic projections and then our budget projections. we had the statements, the labor forces larger by the end of the 10 year window. everyone is flipping on page 50 and 51. it is box 2 -- one. you can see that one up and up on the screen is from page 51, showing the change in the labor force from what we had last year and reflected mainly on the immigration. a couple of things. one is the effect on the economy . that economic impact is the largest impact of the increased labor force. more people, more workers. a larger economy and then from

there, it has budgetary secrets. they have many effects. there are social effects, security effects. many things like that. and of course we are focused on the economic and budgetary. i'm not saying those are the only effects that i recommend. in some sense, some of the biggest fiscal effects might be on those levels of government. state and local government around the country would have physical effects and then effects on discretionary spending since much of that activity is federal government. this response to the immigration search and discretionary spending. we will project that out. on the revenue side, the calculation we did is that those

additional workers would freeze the amount of gdp by 7 trillion and that would correspond to let -- correspond to roughly actually dollars in additional revenue. and then i am so sorry. i think the next piece of the question was the projection of the surge. we had to in 22. it is very difficult to now, sitting here today how long this lasts, how long this immigration search lasts. we don't know. this is the key source of uncertainty in our projections. and then tapering back down and going roughly to our pre-surge rate of population growth. it is something we would follow.

if there were a change in legislation, we would have to do that as well. i think i answered the questions i remembered. >> clearly it is a big theme of the moment, immigration. a well-designed immigration bill from a fiscal perspective. >> it is a challenging question because we steer clear of the normative answer to the congress. we are on our 49th year. we provide the budget and economic analysis. this is well-designed. that is well-designed.

i apologize with that preface. i think we will answer the question. first you see embodied in our work the number of people who also look at the composition of recent immigrants. it is very difficult enough who are the people come across the border now but we do have information on who were the recent immigrants and what are their ages, what are their skills and so on so we have used the information to model the economic and budgetary impact. that is how we came to the conclusion about the change in the labor force. seeing a disproportionate share of immigrants that are of working age. up to age 54.

the prime age that is one and then we look at the education. there is this composition of the workers and we have taken that and said what kind of skills do they bring? what does that mean for immigration and the founding of new businesses? what does it mean for initial wages? and productivity in the u.s.? there is a discussion that initially they will go into sectors of the economy of relatively low productivity but over time their skills rise and they would shift some of that. there is this effect in lowering average productivity and the in turn would translate to wages.

that is the kind of analysis we are doing to come back and connect your question -- what is discussed in the senate now, it changes the composition of immigrants. the cbo would be in position to provide that. >> the growth is faster. because i can go through that again. that is outside of our purview. the discretionary spending, we

know it is there. the surge in immigration has an effect on discretionary spending. the cbo just projects discretionary spending in an -- in a mechanical way. it could be that if the search continues, policy needs to vote on additional discretionary resources and then on the mandatory side, we would analyze that as well, many of the immigrants coming in received work authorization as someone who comes in with parole received work authorization generally around six months or so. they come in with parole and then they could fit in to the

social security system and contribute. eventually they would retire and receive benefits. the benefits side of that would be outside the window. we would have the revenue up front. question is a mix because someone who comes in with parole becomes eligible for work authorization within a year. generally the six month mark. it comes into the labor force. some people come in not through parole but through some other

channel and it just depends on the specifics. also, a demographic report we released in january -- a discussion of immigration. we would track the different channels. this is a term of art that the border patrol uses. someone comes in is not in contact with them. u.s. authorities would generally not receive or confirmation. that gets adjudicated. then they would just be a differential effect on all of this. we would see that we look at people like that and understand that many of them would work just without authorization.

they could look at the different revenue effects of that. revenue effect on state and local revenue -- they were out late for school and sales taxes but the federal income taxes are lower. >> is there any way to engage -- to gauge the percentage of illegals that are in the search? how many of those -- the percentage of people that are illegal? >> we have a breakdown in the chapter. this is in the demographic report from january. it has a breakdown of different categories. in our projections. that has the breakdown. >> caitlin with politico.

you are not commenting on specific legislative proposals. this is why we want to create a fiscal commission. a fiscal 24 government package. you commented on the trajectory where you that debt and deficits going. you can -- if you can comment on that >> i would be glad. it seems to me that the first message -- the projection is a familiar one. that the trajectory is daunting. and you can see in the deficit charts behind me as i flipped to the next one with the debt, forcing the debt ratio -- of course the debt ratio is rising. that is familiar.

on the other hand it is a little bit less bad than it was in our projections last year. that is the effect of the physical responsible of the act of 2023 and the subsequent continuing resolutions that implement to that. it is not enough to solve the problem to remove the fiscal danger but it was enough to be meaningful and show up. that is the way i think of the situation now. we will see if they would support that and if they would work through the budget committees. the budget committee is considering legislation. wherever the congress goes, he would support that. the nc has done that in the past. we will have to get that for you.

when it comes the right time for that story, we will find that picture and get it for you to illustrate. that is what we would do. it could be that our 2020 deficit report might be helpful as policymakers consider options. we do that deficit options report every two years.

>> how much is out there? >> we will go back to john -- it is a really difficult issue. we don't have the real-time information on the backlog of claims and if any claims are withdrawn during the time when the irs halted processing. that is a challenge. >> we increased our projections as we participate under current law. those projections were created effectively in november. we really had no idea what was sitting on those pieces of paper.

we had no idea what the future claims would be. the anticipated there would be much stricter scrutiny going forward. as a result, whether the pipeline is a smaller share. there are more things in the pipeline in november. that is what the irs would be.

>> great. thank you very much. we are here to be helpful. we will come back to you. i owe you a photo. it is a pleasure.