for example, fusion gps is much in the news. they provided opposition research to the dnc against the trump campaign. the data itself and the means by which you analyze and use it is pretty much agnostic, correct? you can use it for or against a political candidate, product, or for them to information campaign, like we said? mr. wylie: yes. said mr.yn: zuckerberg, when he was here the other day, he kept saying we don't sell data. i responded to him, you clearly rent it. i don't know whether that is a fair characterization or not. how would you characterize social media platforms like facebook's use of personal data? they say they don't sell it. how would you characterize it? mr. wylie: they have created a

platform that encourages the use of data. it is true that you cannot go to facebook and simply buy facebook's data. but they make it readily available to its customers via its network of applications or the fact that the layouts of people's profiles on facebook makes it very conducive to scraping data, for example. although facebook would say that they don't allow that, they still create a setup which catalyzes its misuse, in my view. sen. cornyn: mr. zuckerberg also said that the terms of service that consumers agree to when they sign on to facebook, he said people probably don't read it or don't really understand . is your impression that most of the public has no idea about what they are sharing with these

companies? mr. wylie: when you even talk to lawyers who read through the terms and conditions, some of it is even dense for a lawyer. i think it is unreasonable to expect a regular person to have the burden of understanding dense legal texts. the other thing i would say is social media is not really a choice for most people. the internet is not really a choice for most people. it is very difficult to be a functioning member of the workforce or society and refuse to use the internet. i don't know a job that will let you go in and refuse to use google. most hiring requires a linkedin profile. although we use this narrative of choice because somebody is pushing a button, even if they had read the terms and conditions and understood it, they substantively don't really have a choice, because in the modern workforce you have to use social media and the internet. sen. cornyn: they don't really have other options.

mr. wylie: i don't know a job that would hire somebody who refused to use the internet. sen. cornyn: you think it is too much to expect that the social media platforms get consumers informed consent? there is this idea in the law -- for example, if you are going to consent to a surgical procedure by your doctor, your consent must be informed. in other words, you have to understand what you are agreeing to. do think that is too much to ask for in this context? mr. wylie: it is not that it is too much to ask for. people absolutely should have informed consent. but the analogy is not equivalent. when you go and see a doctor and you need a surgery, you need something and you consenting to that surgery is proportionate to the benefit that you are getting. when somebody consents -- quote unquote -- to something online, even if they understand they are consenting to their data being harvested, if that is the only way you can get a job, it is not a genuinely fair situation.

the point that i would make to you is that we should take a step back from this narrative of consent and start to look at the fact that people don't have a choice. they have to use a lot of these platforms to be functional in society and the workforce. sen. cornyn: i don't have use facebook. i can use twitter, i can use -- mr. wylie: but all these platforms do the same thing. as soon as you sign up for twitter, as soon as you sign up to google, as soon as you sign-up to anything, they will be collecting a hugely disproportionate amount of data compared to the utility that they provide to you. further, the other thing that i would say is when you are looking at technology, there is a question of not just informed consent in the present, but reasonable expectations in the future. when i first signed up to facebook, it did not have facial recognition. i posted all my photos, and then facebook develops facial recognition algorithms that can then search the internet and find other things i'm doing. it is not a question of just informed consent. it is about, is it proportionate

to the benefit that the consumer is getting? is it reasonably expected in the future, if there is future developments in that technology, that the consumer did consent to that at the time? and more broadly, this narrative of consent is slightly problematic in the sense that when people have to use these platforms, it doesn't matter whether or not they understand. if they have to use it to get a job, they will still use it. we are sort of coercing and compelling people to hand over a lot of information. which not only now could be dangerous, but you have to imagine in the future, with the developments of technology, what developments we will see moving forward and what kind of risks we will be exposing to people 10 years from now when that data still exists. sen. cornyn: my time is up. i would just conclude by saying that companies now can and do market their services and products based upon people's expectations of privacy. so if consumers are fully informed about what they are doing, what the consequences

are, they can make their choices. it may create markets for other alternative platforms where people's data will be more protected. >> ok. thank you very much for your succint answers, mr. wylie. last month senator kennedy and i introduced a bill to protect consumers' privacy online. included in that bill, giving consumers the right to control their own data by allowing people to opt out of having their data collected and requiring companies to notify consumers of a privacy violation within 72 hours. you previously expressed support of allowing users to opt out of targeting criteria as well as rules to require the data collected by each app be proportionate to the app's actual purpose. with cambridge analytica have been able to harvest the data of of facebook users and their

friends if the users had opted out of having their data tracked by facebook? waswylie: well, the problem that facebook had set up applications that physically allow the collection of friend'' data. one of the other things i would say is that in addition to giving consumers rights, we should be putting obligations on companies themselves. a principle of privacy by design which treats privacy as an engineering problem, as a safety problem, would also be incredibly useful. because privacy is not just a governance issue or terms issue or policy issue, it is a physical engineering issue when it comes to sophomore. sen. klobuchar: do you think they would just do that on their own? mr. wylie: no. sen. klobuchar: thank you, that is my response to one of my friends on the other side of the aisle and why senator kennedy and i have proposed this bill. the honest ads act i have done with senator kennedy and senator

warner, i understand you supported the idea that there should be more transparency of political ads. there is no requirements in place. in your written testimony, you stated, if a foreign actor drops propaganda leaflets by airplane over florida or michigan, that would universally be condemned as a hostile act. but this is happening online. as you know, facebook, twitter, microsoft are now supporting my bill. they are taking measures to disclose these ads, especially facebook. do you think this patchwork of voluntary measures will be the answer? mr. wylie: no. wet in the same way that rea require safety standards in everything else we care about, we should require safety and transparency standards in online applications. sen. klobuchar: in last month's hearing, i asked mr. zuckerberg if facebook had determined of the millions of users whose data was shared with cambridge and let about were concentrated in certain states.

mr. zuckerberg said he would follow up with a state-by-state breakdown of those users. i want to ask if you have any knowledge as to whether those facebook users were any particular states? mr. wylie: i can't say off the top of my head what the density was in each state. i do know there was a particular focus on states from the company's activities on swing states and states that were winnable by republicans. sen. klobuchar: states like wisconsin, michigan, pennsylvania? mr. wylie: yes. sen. klobuchar: we await that information from facebook. mr. riley, i also asked mr. zuckerberg weather roughly any of the 126 million people and may have been shown content from a facebook page associated with a russian troll farm were the same users whose data was shared with cambridge analytica. he replied that he believes it is "entirely possible that there will be a connection there."

what can you say about the potential for any overlap between the facebook users whose data cambridge analytica obtained and the users who were shown content from the internet research agency? mr. wylie: the thing that i would say -- firstly, i didn't ever deal with the internet research agency, so i can't comment specifically on that entity. to your point, my concern is may have beenon shared or misappropriated in the second instance a russian entity from cambridge analytica. is that it isay not just whether or not these individual records were targeted , because if they were used to build an algorithm, whether that was built by cambridge analytica or another entity, other users who share similar profiles and patterns in their data could have also been exposed in a way that if you simply looked at, were these records targeted

specifically, even if the answer is no, that does not mean those records were not used to build a targeting algorithm. sen. klobuchar: i get it. i will follow up on the second round. one last question, if you use the word profiling in this motor disengagement, voter suppression, one of the most horrifying things i saw from our hearing were these ads that were clearly made to suppress the votes of african-americans. do you have any knowledge about the scope of this activity and how often it occurred out of the company? mr. wylie: my knowledge relates to the tail end of my engagement with cambridge analytica. one of the things that did provoke me to leave was the beginnings of discussions about voter disengagement. i have seen documents that reference voter disengagement and i recall conversations that it was intended to focus on african-american voters. sen. klobuchar: can we get those documents? mr. wylie: i will discuss with my lawyers the best way to get

you that information. sen. klobuchar: just to be clear, just as you said, it would be an illegal act to have an airplane come with those pamphlets. this under american law where you are specifically suppressing the vote is also an illegal act. mr. wylie: to be clear, i did not partake myself. sen. klobuchar: no, no, i understand, and that is why you left in part. i get that and i appreciate you coming to testify very much today. >> the chair has asked me -- the chair of the committee and also happened to be in the next order -- i'm not taking chair's prerogative, it is my turn. thank you all for being here. i think this is a very helpful hearing, i intend to be here for the second round as well. mr. wylie, i just leaned over to the ranking member and said she was consummating you on your technical expertise. -- she was complementing you on your technical expertise. i love having a witness before the stand that i can actually

understand. i cannot necessarily understand lawyers but i can understand data analytics people because i was one. i hope the result of this hearing is trying to figure out what if anything congress should do from a regulatory framework with respect to the new avenue for data use. all of us in these committees have already used the data from aggregators. we take the voter data, we know what voting propensities are that are downloaded from the board of elections. you use that as a basis for targeting voters. then there was the next wave, data aggregators, so you can overlay people's affiliations with associations, magazine descriptions. that has become passé in terms of data matching. it has been happening for 10 or 20 years in our campaigns. i would dare to say that every single member here has had that. that is where they buy data and aggregate it and build their proprietary platforms around it. now with the advent of social

entities who have come into play who really don't want to sell their data, they want to sell the analytic result of that data so that they can target people on certain social media platforms. would you agree with that analysis? mr. wylie: yes. it is the age of access rather than the age of transfer in terms of data. sen. tillis: do you believe that data, that baseline, is part of their intellectual capital, part of their institutional value? sen. tillis: yes. data is incredibly powerful -- it is like the new oil. sen. tillis: by the way, i would be happy for either of the other two witnesses to chime in. a part of when mr. zuckerberg was before us some weeks back, i tried to focus on what we should all be looking at as policymakers, as other practices beyond cambridge analytica. i should say that my firm

engaged cambridge analytica. i met them the day that they proposed it and i saw them again on election day. i do have some questions about the focus on trying to get psychographic data from people through in person interviews. it may or may not have been captured through some of the techniques we are discussing today. i also want to go back and talk about, if we are going to do a thorough, impartial, nonpartisan review of the facts, we really should go back probably the last 10 years. i don't know if you had them the opportunity, but i would commend it to you to read -- the m.i.t. technology review had a three-part series, titled how obama's team used big data to rally voters. if you go through that very well written review and through ,rticles associated with it there were quotes from campaign workers who said, we literally took the whole social graph from facebook and were able to

download it through the use of an application. there is a term in here -- i don't know the specific page -- they talk about gameifying the apps that campaign workers would use and possibly the apps that would be supporters were the is. and then the obama campaign, they actually asked whether or not you are a facebook user, if you would mind by clicking on a button, be willing to share information about all your friends. in my case, i have about 4900 friends. so by clicking that button, i was actually giving access to thousands of people information without their knowledge. this was a documented practice that i think also has to be looked at in the context of creating good policy that is based on various uses, contemporary uses of data on social media platforms. in my remaining time in

the first round, i would ask you, do you believe that some of the technology players today simply have grown so large so quickly that some of what you thought would have been captured through good corporate governance and code of conduct, having somebody from a social media platform who either leans conservative or liberal putting their thumb on the scale and giving people information that they really shouldn't if they were good stewards of the data in the social media platform that they were on. i will leave that as an open question to anybody on the panel and reserve my remainder for the next round. >> i will just say that i don't hold a view that facebook is somehow necessary for our lives. as an instructor of political science, i encourage my students never to share information about politics or ads or news online. because they are not professional editors. they should direct their attention to news sources that have editors.

the idea that you must contact your friends and family and upload pictures and that there is no alternative is something that, not as an expert but as an ordinary citizen, as something that is true. in answering your question, facebook has a lot of data. i think it has acted really inappropriately in terms of how it has conveyed news and sold as to anyone. we have about rhythm, we don't pay attention to it, and it is not our fault if it goes wrong. i think it is a terrible way that they have conducted business. i want to encourage people to use facebook, but i am surprised by people's continuing interest in the company. >> i would concur a lot in what he was saying. i think it is difficult to argue that facebook, per se, is necessary for people's lives when about half the internet and therenot on it are lots of other types of social media. to your point about, with

someone with a particular ideological political view tilt the scales one way or another, i don't know if people would deliberately do it -- they might -- but by nature, they will. what i view is true is influenced by how i think ideologically. things that i might screen out and say, this is bad, this is disruptive to the community, would be influenced by how i think. thank you, senator tillis: and thank you to the panel for the chance to be with you today. the reason this matters is we are talking about the intersection for several important developments that are difficult for the average american to understand. big data and social media. and foreign interference in our 2016 election. we are trying to tease out what exactly happened or didn't happen. mr. wylie, let me start with you. cambridge university professor michael kaczynski studied the use of facebook data and found based on average, 68 facebook

likes by user, what you liked or disliked. you could predict sexual orientation, political party affiliation, race, with 85% accuracy. further factors like religious affiliation, alcohol or drug use , even whether your parents were divorced, could be deduced. is that your understanding of that analysis? do you think cambridge analytica and the work of professor cogan used that predictive power to develop algorithms that then weaponized differences between americans for electoral gain? mr. wylie: so the basis of the research we were doing at cambridge analytica was from the papers that you are citing from dr. kaczynski. approachreplicated his and sought to improve it. did i correctly

summarize the incredibly high correlation you could show in terms of really knowing the individual facebook loser -- facebook user if you had access to their likes and friends and activity? mr. wylie: the particular paper that you are citing, the number of likes once you surpass 100 and 200, you can get to the same level of accuracy at predicting, for example, personality traits as your spouse would if they were answering questions about you in comparison to how you would answer. sen. coons: to the point that sh madeor her earlier, all of us who have stood for election have struggled with the difficulty of targeting our voters effectively. the data sets we have had access to, mostly publicly available data, are very thin, very narrow. there is very little that we have. the data sets that facebook has access to, that is why it is a multibillion dollar company. they are unbelievably deep and rich. it is unlike anything we have

had to confront before, correct? mr. wylie: yeah. i don't contest what professor hersh was saying in the sense that -- through persuading somebody compared to motivate them to turn out, much more difficult, and the data sets that traditionally were used are also very sparse and not necessarily reliable. the data sets available now through facebook are orders of magnitude -- and that is why cambridge analytica ended up pursuing that path. it found that in comparison to traditional marketing data sets, the data that you could procure from social networking sites was much more dense and actually much more reliable to create a precise algorithm. sen. coons: to be clear, a billionaire american mega-donor and supporter of the trunk campaign -- and supporter of the trump campaign funded a shell corporation, cambridge

analytica, to take advantage of this research and gained access to 86 million americans facebook information and developed some of the algorithms that came out of that. i want to ask you in the time i have remaining about your experience working with steve badin, one of president trump's senior campaign advisers, and the goals he used cambridge analytica to achieve. was one of his goals to suppress voting or discourage certain individuals from voting? mr. wylie: that was my understanding, yes. sen. coons: was voter suppression a service that u.s. clients could request in their contracts cambridge analytica well ben and was vice president? mr. wylie: yes. sen. coons: so steve bannon is running an organization where you could as a client request in contracts voter suppression using this remarkable data set? mr. wylie: i don't know if it was referenced in contracts, but i have seen documents that make reference to it in relation to client requests and services provided. sen. coons: last question.

you testified that back in 2014, cambridge analytica set up focus groups and pulling on americans views of the leadership of vladimir putin and russia's expansionism into eastern europe. you have also testified that it is entirely possible and what have been relatively easy for russian intelligence to put a keylogger on professor cogan's computer and get access to this full data set. why do you think cambridge analytic was testing putin's aggressive tactics? and what threat would oppose tower 2018 elections if this entire data set, all the algorithms that go with it, are currently in the hands of russian intelligence? mr. wylie: i don't have a clear answer as to why the company was so engaged in testing russian expansionism and the leadership of vladimir putin. that is a question better put to steve bannon. in terms of the dangers for not just american democracy, but other democracies around the

world, this data is powerful and if it is put into the wrong hands, it becomes a weapon. we have to understand that companies like facebook and platforms like facebook or twitter are not just social networking sites. they are opportunities for information warfare. not just by state actors, but also nonstate actors. we really do have to look at protecting cyberspace as a national security issue, in the same way we have agencies to protect our borders, land and sea. thank you very much -- sen. coons: thank you very much, mr. wylie. >> thank you, mr. chairman, and welcome to each of the witnesses. thank you for being here. i think americans are rightly concerned about privacy and the security of our data. we are also concerned about the power that is being collected in silicon valley, a handful of companies having enormous troves of data they are able

to use and employ with very few rules governing what they do is -- what they do with that information. much of the media attention recently has focused on the data operation of the trump campaign, but the trump campaign was hardly the first to employ data in a very significant way in a political campaign. dr. jamison, you talked about the obama campaign in 2008 and 2012 and their data operations. can you share with this committee what the obama campaign did regarding data in 2008 and 2012? mr. jamison: sure. during the 2008 campaign, what they did is they had as an advisor a cofounder of facebook who helped them understand how they could use facebook and obtain data from facebook. a lot of that work was done by consultants, but that was the center of it. in 2012, the campaign changed its strategy.

it pulled it all in house and made it more effective. they were able to combine facebook data more carefully with other sources of data and do a much better job of understanding what individual voters were like, who was connecting with who, and how they can understand those conversations. sen. cruz: and has facebook approached access to data on a evenhanded matter, candidates from whichever party the same access to data? or have they been more political, partisan players in that regard? mr. jamison: i have no knowledge of that. thatcruz: i would note carol davidson, the director of data integration and media analytics for obama for america 2012, she said, facebook was surprised that we were able to suck out the whole social graph , but they did not stop us once they realized that was what we were doing. she also said that facebook

"came to office in the days following election recruiting and they were very candid that they allowed us to do things that they would not have allowed someone else to do because they were on our side." so that is the head of data analytics for the obama campaign, saying facebook was giving preferential treatment to the obama campaign because that is the political side they were on. did they give the romney campaign the same access to data in that election, in 2012? mr. jamison: not to my knowledge. sen. cruz: how about in 2016? there has been a lot of discussion about what the trump campaign did with data. did the hillary clinton campaign have a state operation? mr. jamison: i presume so, but i'm not familiar. there was not very much written about her campaign on any of this. sen. cruz: so does anyone on the panel know what the hillary clinton campaign did on the data

side? does anyone on the panel think that there is a chance in a million years that the hillary clinton campaign did not have a substantial investment in data analytics? mr. wylie, sure. mr. wylie: so campaigns across the spectrum in the united states use data. just to your point, when we look peopleou know, a lot of are concerned about, for example, the ability of big government to inhibit our liberty and choice. big data can engineer a situation that limits our choice and our freedom. it is not a partisan issue. to your point about other parties using types of access on facebook, i actually agree with you in the sense that there is a substantial risk of distorting

the electoral process if a company like facebook decides to pick a side, whether that is democrat or republican. the thing that i would hope this committee and others internalize is that we are talking about cambridge analytica, but it is not a partisan issue. we are talking about the future of how these technology companies operate and the risks to ordinary american citizens and the risks to the integrity howard -- of our democratic processes in the united states and around the world. that is not a partisan issue. >> i very much agree with that and i would note there is an overlay on top of that that facebook and other social media companies are now the vehicle through which some 70% of americans get their political news. censorship, i of think, is a profound threat to liberty, so i appreciate this panel being here. thank you for your testimony.

>> senator blumenthal? sen. blumenthal: thank you, mr. chairman. welcome to you all and thank you for being here. mr. wylie, are you aware of conversations between cambridge analytica executives and any representatives of the russian government or people associated with the russian government? mr. wylie: i'm aware of meetings that the company had with luke oil, which has close connections with the russian government. sen. blumenthal: are those conversations documented anywhere in any letters or emails or any other kinds of evidence? mr. wylie: there are documents pertaining to the conversations and presentation made to luke oil, and i have passed on those documents to the authorities. sen. blumenthal: to this committee?

mr. wylie: not yet to this committee, no. sen. blumenthal: would you be willing to provide them to this committee? mr. wylie: i will consult my lawyers as to the best way to get you the information. sen. blumenthal: thank you. during your time at cambridge analytica, my understanding is you were aware of the founders or funders, including robert mercer, providing moneys so that they would not be "necessarily considered campaign contributions." are you aware of conversations to that effect by mr. mercer or anyone else? mr. wylie: what i am referencing is what was explained to me after i inquired as to the relatively convoluted setup that was happening in the united states with respect to the setup

of cambridge analytica. what was explained to me -- this is not necessarily to say it was the primary reason, it could be an ancillary benefit -- but when you invest money as an investor into a company that you own, it does not necessarily constitute an electoral contribution. sen. blumenthal: me just cut right to the intent of my question. was there a specific explanation to you that the purpose of structuring these funds as investments was to in effect avoid the reporting requirements or other provisions of united states election laws? mr. wylie: it was explained to me as a benefit of the set up. sen. blumenthal: were there any firewalls during the time you worked at cambridge analytica that separated work on different campaigns or where the funds in effect co-mingled in all the

campaigns? mr. wylie: whilst i was there, i did not see firewalls being set up, or any sort of barriers between people or conversations. if that is your question. sen. blumenthal: so there was no recognition of the legal responsibility to separate campaigns. mr. wylie: i am aware of memos from the company's lawyers to some of the executives of the company that outlined the responsibilities of the united states to separate contract between staff and activities that relate to different campaigns were passed, etc. sen. blumenthal: those memos were instructions -- mr. wylie: instructions, or rather advice, from lawyers to executives. when i was there, i did not see -- sen. blumenthal: in practice, the instructions were not followed. mr. wylie: no. sen. blumenthal: can you provide

some specific examples of your direct knowledge of either focus groups or other efforts to suppress voting? mr. wylie: in terms of specifics, i am happy to work with the committee to give a more full explanation. i understand there is limits on time. of research that was being looked at about what motivates and what de-motivates different types of people. if you focus on messaging that de-motivates certain types of people, you decrease the amount of turnout. sen. blumenthal: thank you. dr. jamison, you served as part of the agency landing team in the transition for the federal communications commission for president-elect trump, correct? mr. jamison: yes. sen. blumenthal: during that time, did you have any contact

with michael cohen about fcc policy? during the transition? mr. jamison: no. sen. blumenthal: did you ever meet him? mr. jamison: no, not to my knowledge. sen. blumenthal: thanks, mr. chairman. >> senator harris. sabo harris thank you. -- en. harris: we are here today to talk about how cambridge analytica obtained a sensitive data from millions of americans and used that data to influence our elections. there are broader issues of privacy that are highlighted by this incident. it is worth stepping back to pull all of this into context for the american public. to put it plainly, most americans have entered into a bargain with facebook and other web service providers in which users unknowingly give those companies huge amounts of personal data in exchange for the free service of social networking. in turn, facebook uses this data to show its users carefully targeted ads which are the

source of 90% of facebook revenue. interestingly enough, this business model makes the facebook user product and makes the advertisers the customer. but let's be clear. this arrangement is not always working in the best interest of the american people. first, users have little to no idea just how facebook collects their data, including tracking the users' location, the device they are using, their ip address , and activities on other websites. to be clear, this occurs whether or not they are logged into facebook and whether or not they even use facebook. let me put this in perspective. in the real world, this would be like someone following you every single day as you walk down the street, watching what you do, where you go, for how long, and with whom you are with.

for most people, it would feel like an invasion of privacy and they would call the cops. and yet social network sites technically lay all this out in their terms of service. but let's be honest, if you americans can decipher or understand what this contract means. second, as this hearing illustrates, americans do not have real control over the data collected on them, and there is almost nothing that users can do once data is shared with third parties. i believe it is therefore governments responsibility to help create rules that yield a better and more fair bargain for the american people, one that respects their rights as consumers and their privacy. in the meantime, i have a few questions for the witnesses today. mr. wylie, you have mentioned there has been a lot of discussion about how cambridge analytica targeted african-american voters and discouraged them from

participating in elections when steve bannon was the vice president. what specifically steve bannon or anyone else decide motivates or d motivates african-americans to vote? mr. wylie: it is not just focusing on racial characteristics of people. actually, when you call a -- when you pull a random sample of african-americans, they are not all the same people. they have very different lives and motivators. when you are looking at any program, whether it is motivating or demotivating someone, understanding their internal characteristics is a very powerful thing, because you don't treat them just as a black person. you treat them as who they are. that can be used to encourage people to vote or discourage people to vote. sen. harris: how specifically than did they target african-american voters, understanding as you do that the african-american population is not a monolith? how did they decipher and

determine who was african-american so they would target them in their intent to suppress the vote? mr. wylie: racial characteristics can be modeled. i'm not sure about the studies that my colleague here was referencing. but we were able to get an auc score, which is a way of measuring accuracy for race that was .89, i believe. sen. harris: auc is what? mr. wylie: area under the receiving operations characteristic. it is a way of measuring precision, which means it is very high. to be clear, i didn't participate on any voter suppression program. so i can't comment on the specifics of those programs. i can comment on their existence and i can comment more generally on my understanding of what they were doing, but those questions are better placed for steve bannon. sen. harris: i join in the request for any documents you

can share with us that are evidence of the conduct you describe. on a different note, what should facebook have done to verify that you or cambridge analytica had deleted the unauthorized data? mr. wylie: i can't speak for cambridge analytica, i can speak for myself. sen. harris: please. mr. wylie: in 2016, they sent me a letter that said, we are aware you may still have data from this harvesting program. it informed me that dr. cogan actually did not have permission to use the program he was using for commercial purposes, only academic purposes. that was new information to me, i did not know that at the time. he requested that if i still have the data to delete it and then sign a certification that i no longer have the data. sen. harris: did it require you to sign it under supervision of a notary?

or just sign and send it back? mr. wylie: it did not require a notary or any third of legal signed theso i certification and send it back and they accepted it. >> senator harris, just as housekeeping, we will have a vote called around noon. senate time. we know that means we have to be there within 25 minutes. i think we will be able to get the first round. we have senators durbin, rona, and booker, and i am happy to stick around until the end of the vote if we can agree to three minute rounds. senator durbin. sen. durbin: thanks, mr. chairman, and thanks to the witnesses. senator harris, i thought your presentation on the issue of privacy was spot on. it really describe what we are about, at least one of the things we should be about. i asked a question of mr. zuckerberg, whether he felt comfortable telling me the name of the hotel he stayed in last

night. after a couple seconds of hesitation, he said no. i have asked a lot of questions in the senate. that one god a lot of attention, i think because it really got to the heart of the issue in terms of mr. zuckerberg and his own feeling about personal privacy and where he would draw the line. we know, of course, from what senator harris has said and what isall from life experience, the last hotel i stayed in is probably a matter of some record with my name attached to it somewhere, who knows. but privacy is a critical issue here. and the right of facebook or any entity to use my information without my express permission is over the line. i said to senator coburn, senator cornyn, my friend, when he said he thought consumers were aware when information was being gathered on him, that i'm sure he's wrong. we have now put a little piece of electric tape over the camera

on my laptop. >> good. sen. durbin: most people do now because they are being watched and they may not even know it. but there are two other issues here that i would like to spend a moment addressing. one of them relates to mr. knicks, a british national? is that correct? mr. wylie: yes. sen. durbin: he was clearly involved in some of the campaign activities of cambridge analytica. mr. wylie: he was the ceo of the company, so he was the point person for all of the clients. and he often made the presentations and recommendations to those clients. sen. durbin: the federal election commission says expressly that regulations prohibit foreign nationals from directing, dictating, controlling, and directly or indirectly participating in the decisionmaking process of any person with regard to any election-related activities in the united states. so there is a red flag, or red union jack, whatever you want to

call it, that should make it clear that we're in a territory here that may be -- may be -- a violation of law. you have said that the involvement of cambridge and successor organizations in russia came after you left, is that correct? mr. wylie: sorry, can you clarify slightly what you mean direct asxperience is it relates to lukoil, as it relates to understanding the research that was being done in terms of the focus groups and all that. sen. durbin: in the use of information by the russians in the election campaign -- mr. wylie: oh, i'm sorry. i misunderstood. yes. sen. durbin: that either happened some other place or after you left cambridge analytica? mr. wylie: yes. i was not aware at the time that there was activities in russia to influence the united states elections. sen. durbin: so i would just add

that to the second category. the first question, privacy. the second question is the involvement of foreign nationals in the united states campaign. whether mr. knicks personally or russia as a country tried to influence the impact. the third has been brought up by my colleagues and i think gets to the heart of another very important issue, and that is the issue of the secrecy of this activity. the fact that we know mr. mercer was engaged in this is because of something called open secrets and other sources that were not disclosed in the ordinary course of business in this. did you have any guidance from cambridge when you were there in terms of the secrecy of the clients that you were working for? mr. wylie: so, to your first point about non-u.s. nationals working in u.s. elections, cambridge analytica received formal legal advice, which i disclosed to the media. and that legal advice did make

clear that the company should not be sending non-u.s. nationals to work on american elections. with respect to the secrecy, everybody had to sign a nondisclosure agreement, a very thorough nondisclosure agreement, and in fact after i left, you know, when the company engaged in a protracted legal correspondence with me, it also insured that myself and other people who decided to leave find an undertaking of confidentiality, which is a higher threshold of nda. sen. durbin: you have been asked for some documents earlier, and those two you just referred to, the legal opinion about the involvement of cambridge employees -- mr. wylie: i made it public. sen. durbin: if you would not mind sharing it with the committee, as well as a copy of

the nondisclosure agreement that you signed or others might have signed, that would be helpful, too. >> senator hirono. sen. hirono: thank you. professor hersh, elections very often turn on voter turnout. you would agree that efforts to suppress voter engagement or turnout should be a matter of serious concern to us. mr. hersh: certainly. sen. hirono: you testified that it is easier to demobilize people than it is to mobilize them. you testified on the impact of voter id requirements on voter turnout. if we were to -- consider regulating anything, and i would these -- and i realize i there are privacy isss and it is very complicated. if you want to focus on some sort of regulatory scheme, should we focus on regulating ads or messages that the regulate people? mr. wylie: this is very complicated. i worked as a witness for the obama justice department on government voter suppression, voter id laws. this is really different, and

it's different because campaigns do things that are not nice all the time. if a democratic campaign were to remind trump voters of all the moral failings in trump's past so they don't vote for him, is that demobilization or not? sen. hirono: i know it is not easy to figure out whether a message d mobilizes anybody. but let's assume we can come up with a way to find out what the mobilizes, because that is a matter of serious concern. you said that is easier to affect. that may be an area for us to pursue in terms of any kind of regulation in this very complex space we are in right now. mr. hersh: that might be right. i am not a regulatory expert. in terms of first amendment issues and so forth, if i were to ask someone who they're voting for and they said president trump, and i said don't vote for him, i would really rather you not vote, it does not seem to me that that is a form of voter suppression. even if it is -- sen. hirono: we are talking about people basically paying to

suppress votes. as mr. wylie testified, steve bannon was running an operation where clients could request motor suppression messages which they would then pay for. we're talking about a different kind of circumstance than somebody saying don't vote for so and so. i think you understand the differences. mr. wylie, you obviously have an awareness of the use and misuse of massive amounts of data. i want to ask you this. the u.s. immigration and customs enforcement has proposed a new extreme vetting initiative and their plan is to hire a contractor to exploit publicly available information such as media, blogs, public hearings, conferences, academic websites, social media websites, such as twitter, facebook, linkedin, to extract pertinent information regarding targets, to determine who will be a productive member of society and who will commit crimes and terrorist acts. we are talking about people who are wanting visas to come to our country.

according to the nonpartisan brennan center for justice, ice 's plan would automatically flag people for deportation or visa denial based on the exact criteria from the original muslim ban. it was the original muslim ban of the president that set up this extreme vetting program. so do you think that that kind of prediction of human behavior, as to whether somebody is going to become an upstanding, computing member, or whether that person is likely to become a terrorist or criminal, is that even possible from the information available to the u.s. government? mr. hersh: two points. there is no universal definition of a bad person, and therefore, it's a social construct, and it is laden with people's moral judgments. there is no mathematical way to determine whether someone is a bad person in the abstract, in the sense that there is no

universal definition of a bad person. so the second point that i would make, though, is that just because you are using data science and mathematics, you could have the most advanced neural network made yet, but if the underlying training set for that algorithm uses systemically biased information, so for example, if you are looking at criminal justice statistics and you have a model that ends up predicting that this african-american is more likely to commit a crime because more african-americans end up in prison, you create an algorithm which is simply reflecting social and moral biases of the data sets. sen. hirono: i think the predictive value of what the government is pursuing is very questionable and problematic in terms of all kinds of issues. privacy, you name it.

and the accuracy and ability of that kind of initiative, yet they are pursuing it, by the way. so i think it's very problematic. professor hersh, would you agree? government is getting their hands on all this kind of information so they can determine whether somebody would commit crimes or somebody is going to be an upstanding citizen? does that even make sense to you? mr. hersh: no. sen. hirono:. thank you. >> senator booker. sen. booker: mr. wiley, i have a lot of concerns about how these pitforms can be used to people against each other. one of the greatest ideas of america is the idea of indivisibility, that the lines that tie us together are stronger than the lines that divide us as a country. this morning i listened to the new york times podcast, which was all about sri lanka. the headline was when facebook rumors insight real violence. it was about how facebook was being used by sinister forces to incite hatred between groups,

fear between groups, and ultimately, violence in this circumstance. i was just so deeply disappointed and anchored when nigel oaks, the founder of the scl group, would say things in recorded conversation -- things like this that resonate. to attack the other group and know that you are going to lose them is going to reinforce and resonate your group, which is why hitler attacked the jews. because he did not have a problem with the jews at all, but people do not like the jews, so he just leveraged an artificial enemy. that's exactly what trump did. he leveraged muslims. trump had the balls -- and i mean really the balls -- to say what people wanted to hear. this is to me really a frightening reality, threat to the very ideal of a nation that was not founded because we all but alike and look alike,

we have a set of democratic principles that unite this country. our founders talked about pledging to each other, not religious alliances, not racial alliances, but our sacred honor, . and i am wondering, first and foremost, your experience with this organization that you ultimately left. did you look to see -- did you feel that this was a manipulation of the data to prey upon divisions, prejudices, biases, inflaming them in order to produce, as demagogues often do, a certain electoral outcome? mr. wylie: sure. the united states went through a civil rights movement a couple decades ago in an attempt to desegregate society, and one of the things we are seeing now is a resegregation of society that is catalyzed by algorithms. some people call that echo chambers. an echo chamber is where you start to receive information from one side and stop seeing

information from another side with respect to what cambridge analytica was doing, it was looking to find ways of exploiting certain vulnerabilities in certain subsets of the population to send them information that will then start to remove them from the public forum, if you will, where they start to see more and more conspiracy theories or more and more vitriolic messaging, and they start to internalize that messaging and discount mainstream media. and if there is enough money being spent on targeting these individuals, they will stop seeing any mainstream media online. and from that point, they have now been removed from the public for. and for me, that is deeply problematic because when you look at what democracy is

supposed to be about, there is an element of common experience that voters need to have to collectively make a decision. if you are segregating people in terms of the information space so that one set of voters only see one thing and another set of voters only see one thing, we have destroyed the public forum. senator booker: that was an excellent observation. from your public statements, you talked about leaving as they are creating videos to try and inflict hatred towards muslims, fear of muslims. that's correct, right? mr. wylie: yes, there were videos created by the firm. there is no other way to describe it other than sadistic and islamophobic. senator booker: i have 30 seconds left, but just to shift, my colleague, chairman tillis, talked about obama using big to rally voters, which i think -- i

am somebody that is also very interested in big data and the potential it has, predictive analytics and how it can help from empowering police officers to you name it. we can talk about that but this was stunning to me, to use the data not to rally voters but to suppress voters based on their ethnicity. your testimony -- am i correct in you saying this was a determined effort to suppress not just black votes but the votes of asians and other ethnic minorities. wylie: my understanding is it would target any characteristic that will begin to vote for the democratic party, and in particular african-americans. do you booker: understand where that was coming from? do you have any insight into motivations there? mr. wylie: you would have to ask mr. bannon about his own views. senator booker: thank you very much.

>> i think we have a few more more minutes. senator kennedy would you be , open to a three minute round with a hard gavel with a three-minute mark? senator kennedy. senator kennedy: [indiscernible] dr. jamison, i want to be sure i understand your testimony. in response to answering his -- mr. cruz's question. in the presidential campaign between governor romney and senator obama, now president obama, the facebook cofounder shared facebook information with the obama campaign that it did not share with the romney campaign because the facebook cofounder wanted mr. obama to win. is that accurate? -- wylie >> i did not mean to say that. it was that he was working for

the obama campaign -- a cofounder of facebook. he was working for the obama campaign. i do not have any information i says he took information from facebook, but he explained to the campaign, here is how you use facebook. senator kennedy: did he share that information with mr. romney? dr. jamison: to my knowledge, no. senator kennedy: or the romney campaign? do you know why he shared this information with the obama campaign? dr. jamison: he actually volunteered to work for obama when he was running for u.s. senate, so they had had a long relationship. senator kennedy: so he wanted president obama to become president. dr. jamison: yes. senator kennedy: ok. here is one of the things -- to me, there are two issues. at least two issues regarding the social media platforms. first is the privacy issue.

i happen to believe that social media platforms led by facebook have the ability to influence what we believe, how we vote, what we buy, how we feel about ourselves, and they have a responsibility to disclose that to people and how they do it, and if people still want to use facebook, fine. i think it is a tax on poor people. but people understand that the odds of winning the lotto are not great. i am pretty libertarian. people want to play the lottery, power to them. this is the problem, mr. wylie, that i think we can spend several days talking about it. we can all agree that poison is being spread on social media. here is the tough part. define poison. i don't want facebook censoring what i see.

it does not -- it does violate the first amendment if somebody wants to run an ad on facebook that says don't believe the mainstream media. they are biased. if you tell them they cannot do that, than you do not believe in the first amendment. now, it's a lot different if somebody wants to run an ad that says "go kill every muslim you can find in burma." the way you draw that line is tough. >> i will let you finish. but if you got a great memory, we will take a brief recess. i will let you respond very quickly. we are going to take a brief recess. i'm going to run to and from the chamber. there are at least three members who have asked. we will keep it a three minute round. >> 30 seconds. i think senator booker and senator kennedy's comments get

to my view at the heart of the matter which is that we have a basic human response that we are attracted to provocation and extremism. what the platforms are doing online at the basic level is encouraging that behavior, letting us click and share. we are not drawn to things that are cynically responsible or -- civically responsible or truthful. we are drawn to extremism. to me, that is the hardest part about this. it's not about ads. it's what we want to share and whose job it is, whether it is a cultural, leadership, corporate governmental response , to that very human response to share things that are not necessarily nice for the world or good. senator kennedy: professor, at some point, you have got to trust people. >> we are going to go into a recess that will probably be less than 10 minutes or as fast as i can get to and from the capital. we will stand in recess for about 10 minutes. [gavel bangs]

>> if we can have the committee come back to order? senators are interested in the second round. i have mainly now, because i have to preside, in 20 minutes, if we can keep the rounds to about three minutes, i would appreciate it. or keep it turn for about three minutes. we will start with senator whitehouse. senator whitehouse: thank you. mr. wylie i would like to go to , the brexit campaign and the role that aggregate iq played in it. in the aftermath of that campaign, a person named dominic cummings, a campaign strategist for vote leave, tweeted out that cambridge analytica had a small -- a