let's take a look at what's on the show. keeping control of the computers. europe, it gets behind new rules on artificial intelligence. the european union is trying to lead the world. but will its new laws get the balance right between keeping us safe and allowing businesses to make money from this fast developing technology? i'm going to be discussing all of that with these two. there they are, one of the top tech thinkers from the consultancy giant ernst and young. she tells me that companies are already worried that different parts of the world will have different rules. and the big boss of digital europe, whose members include the biggest names in tech such as amazon, google and microsoft. she helped draft the new rules, and she's going to tell me how she thinks europe can catch up with america on al. and to top it all off, joining me on the show is the woman

behind it all. she's been called the world's most powerful regulator, margrethe vestager of the european commission. she's the driving force behind these world leading regulations. and i'm going to ask herjust how can europe strike a balance between protecting its citizens while spurring on unprecedented innovation? wherever you'rejoining me from around the world. once again, a big hello and a warm welcome to the show. artificial intelligence. it is the technology that everyone can't stop talking about. it's already popping up everywhere, from helping us with our health care to our banking needs. but as companies try and work out how it'll help them to make money, politicians are trying to put new rules in place to keep us all safe. the european union is leading the way. its 27 member states have just given their backing to what they hope is a world leading piece of legislation the artificial intelligence act. and by april it should have the backing of the european parliament, which would mean

it then comes into law. the plan is to try and categorise applications of a.i. into levels of risk. now, at the low end, our ai enabled video games or spam filters. high risk ai systems include those which evaluate credit scores and access to banking services and housing. this is where the focus of strict controls will be at. at the totally unacceptable end of the scale are things like facial recognition systems in real time in public places, something already being used in china. and that is one of the big challenges, to make sure that the rules that govern this fast moving technology in different parts of the world don't make it too complicated for companies to come up with new ways to use it. and one reason is the cost. for example, the eu found if a company wants to launch a single ai product, it could cost them anything from $6000 to $300000. that involves everything from designing the product to having lawyers make sure it's within these new laws.

and there's a lot of money being spent, according to one estimate, more than $230 billion this year, and that will more than double by 2027. once you account for hardware, software and everything else, software and everything else. and artificial intelligence could boost workers productivity and incomes. that's according to the imf, the international monetary fund, in its latest forecasts on the global economy. that economic boost is why french president emmanuel macron is just one political leader who's expressed caution about limiting the possibilities of ai. we can design to regulate much faster and much stronger than our major competitors. but we will regulate things that we will no longer produce or invent. this is never a good idea. and just on last week's show, one of europe's top tech bosses told me of similar concerns. in europe, where we have to be careful is that we are not

overregulating these technologies. i'm all in for regulation, but let's not regulate the technology. let's regulate that... all of these use cases, what we are developing are good for our society, are good for our businesses. they are a framework is needed, but we cannot be too aggressive on regulating the technology first hand. so how does europe's approach compare to the rest of the world, and what does it mean for your business? well, to find out, i've been speaking to one of the top tech thinkers at the global consultancy giant ernst and young. julie tanglin, great to have you on the show. and julie, let me start with this, because i'm wondering, how are these proposed regulations impacting how businesses are developing ai across europe? the regulation is having a tremendous impact, erin, a tremendous impact, aaron, as businesses are really trying to get the benefit of ai, looking at capitalizing on the opportunity. and at the same time, the regulation is putting guardrails up in place to help us do

that in a safe manner. effectively, it's as if we're changing the tyres on a car as it's going, and that's really keeping all of us incredibly busy. julie, how do they compare with, well, say, the united states? you know, if i come back to my car analogy, every car is moving at a different speed. some industries are going faster. if i look at financial services, they're really moving fast and capturing the ai opportunity. if i look across europe, they've been really fast in really understanding the risks that al poses, trying to ensure that regulation helps companies go faster with capturing opportunity, but making them more aware of the risks. and i see other nations and countriesjumping on board, going at different speeds, having different focuses, but all moving on the same direction of travel. do you think they go too far? i mean, will they hamper innovation? well, you're so right, aaron. it's about getting the right balance between having regulation so people

know where the guardrails are, but allowing the companies to also capture the opportunity that's there in the market for the consumers and for their customers and their employees. so regulation needs to be there almost like a handbrake. you need a handbrake on a car when you want to run fast, you need to be able to pull that handbrake. regulation is acting as that handbrake. effectively, if you're using a good handbrake, you can even go faster in the corners. and that's what we want regulation to be. having government and companies work together to get that balance, to temper that. how hard or soft do you pull that handbrake is going to be the most important thing. well, talking of speed, i mean, we all know that this is a technology that's changing at phenomenal speeds, which begs the question, can the regulators, julie, can they actually keep up with it? i think we're going to be ok because the regulation isn't trying to get into the code, but they are trying to make sure that we've got guardrails in terms of ethics and values and frameworks in place,

that companies are really thoughtful about what they're doing at al. if we can keep the regulation at that level, we're going to be a good spot. and julie, here's a question i've been thinking about. i mean, what kind of impact will european regulations have on the rest of the world? because this technology, it doesn't have borders, right? aaron, i think you've hit the nail on the head. this is, i think, the topic that's keeping us up awake at night. i know that. i know that the eu believes that they can lead and regulation across the globe and in fact they've been kind of the first out of the door with their new ai act. and we see countries and nations coming together to think about what they're each going to do. i think what we've got to keep in mind is that business is really functioning on a global scale. and this technology, as you've said, doesn't have borders. so making sure that that regulation actually talks to each other and can be consistent at some level is going to be critical

for companies operating in a global scale. and i think this is the piece that worries some, that we get a completely different view on regulation if you're looking at the us or you're looking at asia pacific or you're looking across europe, and if they're not able to talk to each other, we could get very strange answers. this is the piece that still needs to be ironed out, if i'm honest. and julie, let me end on this other big bosses, other ceos on board with this technology, because, again, i'm wondering, have they had to do crash courses in understanding its use in in order to adopt it? you know, i think the ceos are on board, aaron. you see a huge percentage of ceos looking at how i can further their business. and, yes, many have had to do a crash course. and i will say we're at the very beginning of this fourth generation industrial revolution. we're at the very beginning of this.

i think the use cases that we see for al are there, but there are many more to come as that technology grows and becomes more effective. this is really going to empower our workforce to move without work, for without work flow into leveraging insights from raw data. that's never been possible before. we've organized every company along workflows. it's how we do business. it's how we run processes, how we do governance in the future. with al, we can envision a world that allows us to jump across workflow and leverage right down to insights. that's a revolution. and it's going to take all of us a lot of time to understand this, understand what it means, and expand on those use cases. and i'm looking forward to that journey. lots of new stuff. well, on that point, julie tagline from ernst and young, thanks for your time. really appreciate it. and we'll talk to you soon. aaron, i loved it. thank you so much. great to be with you. and looking forward to more on al. well, with some of the world's

biggest companies, ranging from amazon and apple to microsoft, already rolling out ai into the technology we all use, i decided to catch up with the head of digital europe. it represents the opinions of those tech giants, plus 16,000 other companies to europe's politicians. cecilia, a real pleasure having you on the show. and cecilia, look, look, let me start with this. the basics. what are you make of these regulations? well, it is the eu basically making the first ever ai regulation in the world. and of course, the whole world was watching. it does, hopefully with time, create the legal clarity for companies to to implement and to use ai much more actively. but it certainly does also pose many challenges for european companies that are quite far behind.

i mean, only 8% of companies in eu is actually using ai, whereas for example in the us that number is 50% and in china it's 70%. so it will actually give us extra burden and we'll have to see if we manage to strike that balance between a very swift legal clarity and ease of implementation and actually boosting the uptake of ai. well, the question has to be what can be done to to boost the ai uptake in europe? i think, first of all, we need to create that legal clarity. and are you compliant? are they in high risk, which by law that you need an ai sandbox in each country? i hope that they're going to work across country to together so that the countries can basically scale across borders. and last but not least, of course, i mean, venture capital. we need to discuss how we attract more more investment in europe. it's a huge issue. and then more talent again. so talent, i would say we need the competencies, we need the capital, and we need a common market that is scalable across border. and cecilia, do you think they've

got the the balance right between protecting people versus stopping the potential of the technology? i mean, we do support the risk based approach. and actually, we've been a part of course, industry have been consulted and we've been a part of making the risk based approach where we look at the use of the ai, not the technology itself, where we have the risk based approach, where, for example, ai in health care or ai and critical infrastructure. in recruitment of people is considered high risk scenarios where really big systemic mistakes can happen. whereas, for example, your algorithm selecting your next netflix movie is not considered high risk. so i think that actually strikes a very good balance. now, i mean, if you look at the algorithms of foundation models of what we call generative

ai, we were definitely not in favour of having that regulated. it is kind of looking at a specific technology that will change extremely fast and where europe already is far behind. so there will be extra burden on european companies to basically use these generative ai foundation models now, which is quite challenging and maybe not the right way that we would have liked to see this regulated. and cecilia, i want to talk about cost. how much of a cost burden do these regulations place on a business wanting to, well, offer an ai product? what we do know is just indirect costs. you know, hours spent by employees for a company that is like a medium sized or even small company, like 50, 100 people would be around ?300,000. and remember, this is only one regulation. then we have the ai act, and many, many other regulations coming that way. so should we actually create some

kind of grand or allowances or schemes for companies that wants to invest in being compliant and the adoption of al to accelerate this kind of both the compliance but also the uptake and the adoption of ai into small and medium sized companies in europe. we definitely need some kind of accelerated so we don't get even further slowed down. do you think that will mean small businesses are just going to be shut out? yes, for sure. and of course, i mean, you cannot solve everything with regulation. it's a part of the equation. if you want things to grow and to succeed, there needs to be what i said before, you know, competencies, capital and a common big common market. and cecilia, as we know, you speak for tech businesses in europe, at the highest level. so i want your personal view about al's potential impact on thejob market. it's one of the big ai talking points. are people right to be worried?

well, they are right to be worried, but also to be very optimistic. i mean, each time we have a technological revolution, i mean, there are jobs that will disappear and they are new, i mean, there are jobs that will disappear and there are new that will occur. the question is, are we ready to reskill the people into those jobs? and is our schooling system ready to basically know already from childhood, teach them how does algorithms work? you know, how do we work with data analysis, etc. ? and that will be increasingly important because not only do we need them to have the newjobs, but we also need them to live and be voters in a new digital era where algorithms and data and with digital is a fact of life and actually has been for the last 20 years. and cecilia, let me end on this...

five years time... do you think europe will be better off for moving first? i think that it has created a awareness around ai that now needs to be followed by investment, understanding of how to bring the educational system, the workforce into there, and also how to build a digital defense, which is still to be seen. the focus is there, but we need to move faster than we have done before. well, on that point, the director general of digital europe, a real pleasure having you with us. thanks for your time. good luck with everything and i'll talk to you soon. thank you so much. thank you for bringing me, aaron. so as we know, the european union is setting the pace, or so it hopes, with rules about how artificial intelligence can be used. the uk, australia and japan, they're amongst other countries already working on their own new laws, but none are as far ahead as the eu. so i decided to catch up with the woman leading that charge.

margarita vesta, an absolute delight having you on the show. and margarita, let me start with this, because the act is there, but it's not law yet and it's still a long way from being enacted in your member states. so i need to ask you, when are you going to get this across the finishing line? it's a great pleasure to be with you, because this is an issue for everyone, because artificial intelligence is already everywhere. so the law was just approved by member states quite recently, and soon the parliament, the european parliament will vote upon it and then it will come into effect in in sort of in different portions. so the first thing to come into effect would be the prohibitions that you cannot do sort of social scoring. you cannot put artificial intelligence into a doll that will try to trick children into do some things. and then eventually the more sort of technically tricky things, they will come into effect later.

you know i'm going to press you on this. i mean, what's later? i think the last provision to come into effect would be two years from the publication in the 0fficialjournal, which will happen this spring. what have been the main sticking points in getting these regulations agreed to? two, primarily. one, that we thought of from the very beginning, which is how to make sure that regulation and innovation goes together. because part of the point of having legislation and artificial intelligence is to make sure that we feel safe in using it. and because of that, we will press through with innovation, for instance, using it in the public sector, in the health sector. when you distribute social benefits, when people are accepted to a university or another educational institution. and the second sticking point was, of course, how to deal

with large language models, how to deal with general purpose ai. because we first set pen to paper almost four years ago, and back then there was no chatbot. then there was no chatgpt. but the main architecture is the same through all our discussions. we do not want to regulate technology. we want to regulate the use of technology. if the use is risky, you would have obligations if the use is absolutely trivial, like a recommender system for the next song you'll hear on your streaming platform, then go, go, go! are there concerns about the costs i mean, the costs of complying with these regulations? because some will say that the high costs are going to hit smaller businesses more and keep them from competing with the bigger companies who can well simply afford to meet the costs of these regulations. we sort of have an asymmetrical approach. if you're the big guy,

if you have market power, if you can have systemic implications, you also have big responsibility. if you're the small player just trying to make it in the marketplace, we will enable you. we will make sure that the marketplace is open for you. we make sure that you don't have to do the same heavy due diligence as the big guy who might have systemic implication. and i think that is really important because there are so many small players in this field and what should not sort of underestimate the innovative power that comes from smaller players. another potential problem for europe is that it's not traditionally been whether the most powerful computer chips are made that you need for al. so how do you make sure that europe has enough of those? and it isn'tjust regulating tech that's that's made elsewhere?

of of ai is also a risk to your economic security. so we have really been on it both to increase the level of production in europe before the pandemic. i think it was, give or take, 10% of world chips that were produced in europe. we want to increase that to 20. working with the americans who want to go from 20% to 30%, very often you would need a lot of very different chips to produce sort of the full suite that you would need in order to have the full ai chip really working. but i think with europe being a much stronger player and having some of the most essential companies for chips production in europe, we can to a different degree be a heavy weight in that ecosystem

without being sort of overly brutal or or thinking too much of ourselves. we know what the french president, emmanuel macron, thinks. so i've got to ask you this. how do you get that balance right to to make sure that you don't overregulate al or other tech in such a way that european consumers and businesses and the economy as a whole aren't left at a disadvantage to other parts of the world because, well, companies, let's say, in asia or the united states, for example, come up with innovations that you don't allow. well, this is a question that has been with us from the very first day. the amazing thing with artificial intelligence is what you can do with it once you trust it. so the regulation should produce the necessary trust if there is a risk to ai so that you were willing to to go ahead and use it as much as possible. and the first step in doing that is that you can innovate, you can do your research, and you will not be touched by the ai act.

and if your use case is a use case where there's no risk, then you just go ahead. and when you then have some risky use cases that you're actually really careful to make sure that businesses have a fair chance of meeting their requirements. and those are the balances that we have been pushing. well, i mean, you talk about i developing jobs. i mean, something i've done entire shows on. but the flip side of that and many experts will tell us that al is going to cause the loss of many, manyjobs across, well, pretty much every industry on the planet. no, i also think that thatjobs will change dramatically because you have access to a completely different helper or intelligence, whatever you'd call it. but i think it's really important to stay true to some of the fundamentals, which is that technology should serve humans, that al can help you brainstorm, it can support you

in decision making. but i think that you would still prefer that it's your doctor taking a decision than the machine, but you would prefer the doctor to have the best data available before that decision is taken. but do you think ai will cause more job losses than it creates? i think the net effects may be positive, but i think it's really important to be careful in the process, which means that you need to focus on digital skills, on lifelong learning, on how to develop the transformation system so that people can also follow and say, okay, this is not dangerous. i can manage that. it will enable me to do myjob in a better way than what i did previously. and let me end on this. in five years time, what does the ai picture look like across europe? you'll find it everywhere

because you will find it when you're trying to fight climate change to to interpret data. you will find it in your electricity grid to make sure that all the different ways of producing electricity actually matches the demand. when we're cooking for dinner or getting up in the morning, you'd find it. when you go shopping, you'll definitely find it in your health system for a better interpretation of your blood samples and all the different variables to make sure that also very rare diseases actually can be discovered and cured. so you'd find it everywhere. but as we use more artificial intelligence, also take the occasion to renew how we relate to one another. because the most fundamental strategic choice we've made is to make sure that technology serve humans. that is what we are striving for.

well, on that point, a real pleasure having you on the show. good luck with the regulations and we'll check in with you soon. thank you very much. you're welcome in five years time. well, that's it for this week's show. i hope you enjoyed it. don't forget, you can keep up with the latest on the global economy on the bbc news website and the smartphone app. of course, you can also follow me on x. our thanks for watching. hello, there. there were some spells of sunshine around on saturday, but hopefully a good bit more of it to come on sunday. but also some outbreaks of rain. so a bit of a mixed picture and beyond the weekend, it's a fairly mixed picture into next week with low pressure systems tending to dominate and after monday, temperatures actually tending to climb.

but let's look at all of that in detail. first of all, through the overnight period, we can see we've got a fairly wet weather towards parts of eastern england, in towards eastern scotland, elsewhere, a little bit drier with some clear skies. where it's clear we could see some mist and fog and still a number of showers pushing in towards the west and the south. that wet weather, though, across parts of the north east of england in towards eastern scotland, withers first thing on sunday and tracking its way northwards into the northern isles. elsewhere, though, sunday is an improving picture with some decent spells of sunshine and also a few showers coming in on a westerly wind. so by mid afternoon, much of that heavy rain has cleared, although lingering across parts of the north east of aberdeenshire and in towards the northern isles. elsewhere, though, avoid the showers and you'll see some sunshine in temperatures up to around ten or 11 degrees. and then through sunday evening and sunday night, we start to see largely dry and clear conditions and that means it will be colder than overnight saturday to sunday. so this is tomorrow night. sunday night. you can see temperatures there in towns and cities, of frost in the countryside. to monday, a low pressure situated to the northwest.

and we're dragging in some rather cool air from the northwest. so a showers pushing in through parts of scotland could well be wintry on the higher ground. a number of showers in towards northern ireland as well. but elsewhere for much of england and wales, it's a dry and fine day. it will be a chilly start. a touch of frost here and there. but by afternoon, temperatures up to around nine or ten celsius cooler further north. and then from tuesday onwards, we start to see a shift because the wind is changing direction to more of a southwesterly. so we're dragging in cloudier skies and some outbreaks of rain. but the temperatures will be on the rise on tuesday. it should still be largely dry and sunny across scotland, some showers in the north, but the general trend from tuesday onwards is for cloudier conditions with outbreaks of rain. but you can see on the outlook there, the temperatures will begin to climb and certainly for some reaching the low to mid teens. that's the forecast for now.

live from washington, this is bbc news. animalfeed and rice — gaza residents resort to desperate measures for food, as the un warns of a looming famine. a missing six—year—old girl is found dead in a bombed—out car in gaza city, days after her pleas for help,

drew international attention. and republican rivals for president, donald trump and nikki haley, hit the campaign trail in south carolina, two weeks out from the state's critical primary. hello, i'm carl nasman. we begin in gaza, where the humanitarian crisis is deepening. residents tell the bbc they are surviving on animal feed and rice as food stocks dwindle. the un says 300,000 people living in the isolated north of gaza are largely cut off from aid. more than half the aid missions to northern gaza were denied access last month. that's according to the un's umanitarian coordination agency. mahmoud shalabi from medical aid for palestinians says the situation there is dire. when it comes to the flour, people are actually using animalfeed.

however, even the animal feed is starting to get scarce