bridge watch the house transportation infrastructure committee live at 10 a.m. eastern on c-span, c-span now our free mobile app or online at c-span.org. c-span has been delivering unfiltered congressional coverage for 45 years. here's a highlight from a key moment. >> stay with you, and chancellor, the slack, the symbol of -- in in war we stand, we fight and we will win because we are united. ukraine, america, and the entire war.

[applause] c-span, powered by cable. c-span is your unfiltered view of government funded by these television companies and including comcast. are you thinking this is just a community center? no. it is way more than that. comcast is partnering with a thousand community centers to create wi-fi enable lift so students from low income families get tools they need to be ready for anything. comcast supports c-span as a public service giving you front row seat to democracy. >> this is neil of the abundant institute of a.i. policy and chief technologist of the

federal trade commission to artificial intelligence and campaign 2024. good morning. >> thanks for having me. >> let's start with abundant it?itute what is >> a new mission driven nonprofit with our explicit goal is to create the cultural and policy environment that allows emerging technologies including a.i. to germinate to bloom, and to flourish and so we think that is the major road to widespread human prosperity and we're excited to tackle the big problems that -- that we think are in the -- in the united states right now. >> so much so they made you a director of policy specifically to that. how much does that reflect the concern they have? >> a.i. we think has huge potential to create widespread human abundance there's this new set of tools a.i. has been around for a long time and you and i have talked about it in the past. but this new set of generative

a.i. creating new set of tools that i think empowers individuals to really create new and powerful types of communications with each other and so, we think it is one of the bignk tools and then as a.i. continues to be used throughout the economy, we think it is going to create a lot of opportunities for growth, economic growth, also personal growth and we're really excited about it. >> before this job or at least in your resumé you were at the federal trade commission called chief technologist what was that? >> federal trade commission is the general consumer protection, and one of the competition agents in the united states and the chief technologist job is to bring understanding of technology to that policy space and so, i've spent a lot of time look at reports. but also cases, and investigations that the ftc was diagnose to bring that lens of technical expertise into the policy work that the federal trade commission was doing.

>> right before we go into the topic at hand abundance institute how was it funded and financially supported? >> so a wide range of foundation individual, and corporate donors we're new so some of that is still being figured out. but you know we just launched in april but that's sort of the space that we live in. >> if i asked does a.i. money come from the community? >> i'm not 100% sure on what companies but i think that a lot of a.i. communities certain parts of it would be interested in what we're doing. yes. >> so when it comes to topic of the election, your institute put it out as part first of a series of reports forthcoming about the election what's the concern overall at least the thing you're concerned about with these reports? >> so, you know almost as soon as chapter chat gpt launched and there's concern about the 2024 election there's a lot of

speculation and wasn't clear what was happening. there were some reports of some incidents that came about and what weer decided to do was lauh a -- observatory a tracker that would track all media mentions in u.s. outlets of a.i., a.i. use in election and so -- we've been running that using the keyword tool that tracks all of that. and we launched it in january 21st about a year before inauguration, and we've released our first periodic report last week. yeah, 180 days out from the election and so -- you know, out of the 7500 articles that are tracker identified, we spotted three incidents of uses of a.i. in elections. u and some of those are very interesting we can talk about what specific instances are. and so the goal here is to track all of the way up and through the election so after the election we have a very good

ground truth, a database that says like what actually happened? because a lot of speculation is that a.i. will fundamentally transform our elections we've seen a lot of concern from academia from politicians, saying things like, this is our last human election. and/or -- that we need to get this under control or, you know, we won't have a democracy anymore. and so we wanted to like really get a good database of information that says, what kind of impact is this actually having right now? >> before we look at examples to what degree do you think a.i. has ability to undermine or sway election? >> i think there's a couple of different vectors that can happen c one of them i think people were worried about that it would be used to hack election machines. itha turns out that that vector from all of the people i've talked to is very unlikely. the election system, the mechanics of the election system

in the u.s. are very robust and districted so it is a kind of -- a tax surface that's difficult to get around and a.i. doesn't change that much. the other way it might change things would be around disinformation or misinformation. now, the question is how much does it increase misinformation or disinformation and election? i don't think that any watcher of news or observer of u.s. politics is unaware that there's actually a lot of misinformation in election communications already. and so people are generally quite skeptical of information and so, i think on the margins it will increase the amount of speech because these tools do make it easier for people individuals, including individuals to create sophisticated con tngt but there's no reason to think that speechte will -- like be out of the balance more lies than truth. ii think a.i. has a lot of potential to actually help people who have, you know, who are marginalized communities who want to speak to politicians or politicians want to speak to

them, to connect them through things like automated translation so i think there's ways ways that a.i. could help increase robustness of democracy there's ways that it could threaten it so we're doing this research to at the end of the election cycle we'll have a really good set of data that says what was the -- practical impact? >> this is neil joining us abundance institute and questions about artificial intelligence and upcoming election call the line. if you live in the eastern and central time zones, it is mo202-248-8003. from the first(2 part of the report one thing that came bag the information you called together was a robo call that took place before the new hampshire primary. >> this call went from 25,000 people in new hampshire. it is not quite clear how they

were targeted democrats or reallies or some mix. and this robo call was the voice of joe biden saying, if you vote in today's primary the republicans will win or something like that. and so, the -- the implication was that it was trying to suppress voter turnout. and that robo call went out, you know, i think hundreds of thousands of people voted in that election. and the secretary of state hasnd since of new hampshire has since investigated this andth said the was no evidence that it suppressed votes, in fact, the republican primary that year which iss on the same day had a record --ub record turnout that democrat one had a very average turnout for -- a year in which the democratic candidate is running unopposed so it doesn't appear to have a suppressivesu effect and created by a democratic strategist. he says to promote idea and the concern around a.i., and so --

so it is interesting case. >> we have that for the folks here's that ad that took place before the new hampshire primary.d >>e new jersey presidential preference primary republicans have been trying to push nonpartisan and democratic volters to participate in their primary. what a bunch of malarkey. we know value of voting democracy when our vote counts and save your vote for the november election we'll need your help in electing democrats up and down the ticket. voting this tuesday only enables the republicans in their quest to elect donald trump again. your vote makes a difference in november not this tuesday. if you would like to be removed from future calls, please press number 2 right now. ....

it can be used to lie. there are some concerns about that. and so i think the call itself, so we found out did not have a huge impact. but in the something borders should be aware of if they're getting unusual calls that people are selling a certain person they should pay attention and double check. especially their local election officials about what is going on. courses take a call from mika from north carolina. your aunt but the abundance institute formally of the federal trade commission. us today to talk about the 2024 election. whenen it comes to election matters. mika, go ahead. are you there?

okay i think i'm a little transmission probably cannot hear. we will try that call. mika, if you are on keep going or we will get back to in a second period when it comes to this idea that you found about about the video being one part of those concerns of information of the things that you find that came back in the search you did? >> two of the other incidents otwere an incident where there d been a deep fake created of taylor swift of her holding up a sign that says trump a one, biden lost their two different insensible is a video, one was a picture. it turns out it appears especially the photo one could have been down with traditional editing tools and so it's not clear at all that was eight i create a problem. that was on the incidents i got reported as potentially ai problem. the other when we found in the period that we looked at their

work trump supporters who spread ai generated photos of donald trump embracing african-american women and showing support from the african-american community. this was a deep fake. it was created by ai. there's a call in show guy in florida who admitted he use the tools to make that. those of the two other ones within the window. we did also talk about the others that happened before the tracker. that was used in a desantis add of some fake pictures of trump and fauci together. that had happened before that did using ai tool that was in a campaign ad. that is the only would we have seen so far of the use of ai technology in such an explicit way in a campaign ad. >> a highlight in the story

during the ad it was the pictures of where the president presentappears to be hugging dri that has been created. >> i think that is right. slipped in and amongst other pictures as well that were real. you would think this to be the certain thing that would be very hard to notice. because his or high-profile candidates, obviously there's a lot of scrutiny on the ads that they put out. it was identified a relatively quickly this had fake images in it. e erthe desantis campaign was ry certain columns branch of the desantis campaign. obviously the desantis campaign is no long-term not shorts tragedy benefited themth as the way they hoped it would. again another incident in the tracker of 7005 at meter reports we saw this as one of the three or four the keeps popping up a box here is the photo of the bbc publishing that former president with african-american

supporters. esteemed as false the bbc picking that up as well for this one of the conversations we will have when it comes to 2024 matters in artificial intelligence. illustrate mika again and north carolina, good morning. >> caller: i'm currently set it studying artificial intelligence programs in college their major issues i see with it that could affect our political landscape. mostly it is a fact that our political leaders and uneducated on ai. they have no idea how it operates how it's created the training that goes into it. the data sets that are used for. and so on. has your institute looked into trying to create training programs for our congress said that they understand the concept of this technology a little better when they are making policies? i do not want another ai winter to happen as soon as i get my degree. >> i do not want that for you either.

as a call back is interesting. it'ser come to several periods there's tons of excitement and it sort of dies off when it gets integrated into technology nobody thanks of it as ai anymore. they think of it as computers. i think your point is really important. getting a regulation right in in the space is critical. you don't need to know everything about technology but you need to know i like to call the policy relevant characteristics ofle ai. my institute has great educational material. we largely use them at the state level for state legislators were a ton of the legislative action is happening. congress moves slower. it has a lot of competing other interests. and so we are making those inroads in congress as well. they're also been talk to a lot from people from industry and for advocacy groups that are all trying to bring congress up to speed on this stuff.

each with their own politicalan angle. we do think it is important to principal educational materials to help make the right decisions about how to regulate in the space absolutely. stu and eric is in antioch, california hello. >> caller: good morning everyone. i was just wondering how the election is going to come down to five or six states and out of those states i wondering did ai possibly be telling what individuals you have to concentrate on ands be able to tell by algorithms how to influence those individuals to make them vote the way you want? >> it is a great question. political strategists have for decades at this point been selling tools that can help identify it's really hard to change people's mind.

trying to figure out how to do that, ai does not have a particular advantage on that. most of the techniques we use in targeted political advertising techniques that use computers and they use algorithms there are now at the newtonut generative ai does not create a new set of tools for identifying people who are particularly persuadable. i don't want that is a big difference customized messages the generative ai to customize messages may beat to the issues the politician and the targeted voter intersects they most care about. maybe there are tools to help politicians reach their candidates they would not otherwise connect with.

under this new wave of generative ai. >> how much background and technology do you have to used to use ai sophisticated like what we sell with biden. and how much trading do you really need? >> not a ton. it turns out the person who carried the biden a rowboat call was a newll orleans street performer visit democrat strategist had hired to do this. i don't think he had any formal technology training. a lot of these tools are point and click. they are really powerful and interesting. i encourage the listeners to try out some of these tools and become familiar with what they can do and what they can't do. that familiarity with the tools will help people understand theirth limits. and also pay attention to what theyin are hearing and may be being eight bring a skeptical eye to it. with some educated background progress you see a future with some type of transparency is

needed when you look for an ad and see if ai is being used in the ad that needs be identified in some way? >> are a lot of bills looking at that. in the political space there are some bills actively looking at that. we have disclosures in political ads for various things. the challenge is, as i mentioned before once i ate works we don't call it ai anymore. your cell phone camera has ai algorithms built into it. if you take a video with that does that count as an ai created ad? there is really difficult definitional questions about what would count as ai. the safe advice would be too safe i am the lawyer for a candidate i would say put this disclosure and because we are not really sure if you have that disclosure and every addict doesn't add anything. it does not add any information. i think there are real trade-offs it there's important messages the kid has moment to get out to the public.

i think there are real trade-offs and having forced foe disclosure in thatt space. it's certainly something congress is lookingin out. neil chilson of the abundance institute letter from dave in south carolina, hello dave. >> good morning pedro. listen, why is a basic deceit that's being practiced byha ai purveyors, why is this deceit not prosecutable? this type of interference in our election and the control of our government is unheard of. i don't understand why those that have this are allowed to practice what they do thank you. >> it's really great point. especially in the commercial context is already criminal prep you're lying about your products or prices or anything it is criminal. in the political state,

political speeches highly protected by the first amendment. miss truth and it's often quite subjective. however, there are laws of theof westate and federal level againt lying about such things as when the election is or where your polling places. does not matter what tool you used to do that whether use pen and paper sophisticatedte ai viewer lying people to suppress voters and suppress the turnout by lying about the process of the election, that is a crime and it can be prosecuted. >> we have a viewer off of x asking this question. it see i china or russia decides to fully pull the trigger on ai the rest of the western world but have not but have to do the same. you agree with that? >> i'm not sure if this is in the context of the election? >> their ability to influence election. let's take that progressively take it in the elections, again

it's hard to know exactly how ai would be used. and many cases the goals of foreign adversaries and elections is not to determine a specific outcome. uoften it is to undermine the american trust and electoral process and for that you do not need sophisticated techniques. in fact the sketch year the techniques look may be the better they trigger someone's alarm and say it looks like something is going on sketch in the selection. i do think it is not amp up the ability to undermine our confidence in elections. if that is theon goal of many of the third-partyoa actors ai reay does not change that too much because their way to measure how much more sophisticated ai now versus when it first made the scene? looks lots ofti benchmarks these

companies used to test their models against. lots of different things. many different capabilities and i'll be so they are different for video, or for audio or for tax which is three main areas ofgenerative ai right now. what we are seeing as it's getting much much better. chatgpt launched a new version earlier this week. it's very impressive it can do things like live translation between two people. and has a lot of potential meant three, four, five years ago. >> chris in illinois you are on. >> thank you for taking my call. i've a question and a comment. regarding the gas statements earlier in the broadcast when he

said this tool had the capability to increase speech. i was just wondering if you could commentco a little bit of what he meant by that. and then the other question is with regards to regulation, we already note technology typically leads everything and it seems like regulation of follows at a very slow pace behind the speed of technology. things like the example where this is not very technical a great gp add. when i was in the real professionals do it? especially campaign space. what kinds of things as he anticipate could go wrong? it went great set of questions. that's quite related. when i was talking about increasing speech what i meant is it empowers people who do not have sophisticated studios or cameras or editing software to

generate something that is persuasive and compelling and coprofessional looking. that is a level setter. i do think what that mean is people whose voices have been more marginalized or upstart candidates were trying to take on a big incumbent it levels the playing field a little bit between those two. that's what i meant by increase speech. we look at how these tools are being used right now especially the image once just the explosion of rich interesting images is an example of how much these the ability of someone to themselves that holds true in the political space as well. i expect to see contents around politics people are using these tools to speak. you're saying i don't know about all these people have big new

powerful tools that might be a threat to those of us who are in a sophisticated and are tools. ultimately is a good thing for democracy. that is what we want. we want people to express themselves and get their ideas out there. the second part of the question is more about level setter than a level up the job space what we are learning from generative ai it's helping people who are new to jobs who are who are in inexperience very quickly move up the learning curve to become people who are much more than puttingph up much more sophisticated work product in a much shorter time. even if they are not reaching the peaks of the top performers immediately. it does seem to be a level setting process that probably true in the political speech states as wellal regardso

027,487,000 202-748-8001 for the mountain and pacific time zones. if you talked about legislation efforts in congress there's a story today from senator schumer is going to unveil what's being described as a roadmap when it comes to looking through ai the story highlights it's nott legislation the senate task with the roadmap should campaign but would you say? >> the culmination has been having a series of hearings on the hill closed-door sessions. lots of people from industry and talk about the technology and where the issues. the most important piece is how does government use this nttechnology? not just have these in criminal justice or overall administration of government.

how did that make sure the government and military or taken advantage of the technology that is both safe that it respects the rights of citizens and increases the capability of government to deliver it what it supposed to do for the american citizens. that's a really important thing. i is not require pushing a lot of mandates externally. direction from congress what agencies should be doing and thinking about as they are evaluating these tools. >> is supposed there's a sweet spot of regulation when it comes to voting what it does not doesg on the industry overall. >> when it comes to ai because the general purpose like electricity to look at the sectors in which is being used. so for example using ai and healthcare. one of the places it has

enormous potential for customized drugs, customized treatment raised on your specific health needs we already have a pretty robust regulatory regime there. how to say i fit into that robust regime and let's it regulate as it's being used in healthcare as is being used in transportation. as it is being used in telecommunication. rather than trying to attempt to regulate ai generally, which would be sort of like saying let's regulate computers generally. it's a very difficult task it probably will not achieve the goals that were trying to reach for. >> has been part of these discussions when it comes to the roadmap or oversight? >> the schumer hearings were pretty new as i said. a lot of our partners and people that we work with have participated in those. as a schumer roadmap rules out and congressionalou committees e

will certainly be active in educating legislators a writ of what are the right pathways here. >> klobuchar on tech related issues as well. especially in this space. >> klobuchar is due on wednesday to introduce two bills run ai and elections i had testified te her in the previous round and i'm somewhat skeptical of those bills for some reason we talked but one of the mandates of disclosure and transparency. it is overbroad it tackles not just that is political ads for the other one would ban the use of deep faker ai generally technology that has a challenge of how do you define ai? does that include ai on my camera it also is another challenge is that it is broader than political ads. includes any political speech.

if you tweet ai generate picture of donald trump that might fall under this particular law that raises a bunch of first amendment questions given political speech is important and protected by the first amendment. >> was hear from raul in california. thank you for waiting good morning. >> i want to say ai during this time right before an election is very, very dangerous. former trump the side ai cheated him out of the presidency? he already almost turned out the white house once. who would be responsible for that? >> i share your concern. and in this sense. it's less wit ai, our study it shows ai thus far has not had an

enormous impact on elections. that can change and that's why we are monitoring this. but we help is at the end of this process and say whether or not ai on various races. that would directly address the challenge. the second there is a segment of the population that's quite skeptical i do worry if we talk about over and over the threat ai has here the tracker picked up 75004 or three incidents that did not havets impact on the election. that hype and concern could itself lead to bebe skeptical after words of an electoral outcome. i doou worry about that. it is really important as immediate consumers that media

producers out there to be thoughtful in how they talk about the use of ai in particular not jumping to conclusions something was ai generated. many of the video and picture editing tools we have right now are plentyig powerful to create some suspicious photo or video. not jumping to ai is to blame. also following the story through what impact did it have on voters? those two things could go a long waysay towards having people be realistic about what impact ai might have. it will not stop politicians who have outcomes it did not go in there favor by blaming x, y, z for the reason it did not go in his favor voters moderate their own concern.

more realistic eyes what was going on. >> this is the viewer and new iw mexico saying is driven ai programs the theme technology changes the political value today. >> that is interesting. i think often we o talk about ai congress earlyci on we talk abot ai in speech we blend them in a social media context that we are in. a lot of the problems people are saying are ai problems are kind of the same concern about social media primarily. this is a new way to create contents. at the social media or traditional mediari function. when you separate those two that

is it ai causing the problem is that the distribution. it is a concern we have a long hat around distribution of content and different ways. that might be the proper place to focus attention because you say platforms have to take more responsibility and if that is the case how do you keep watch over that? >> depends how you adjust the problem. platforms have a responsibility to create an environment their users want to participate in. they do that and lots of different ways to content moderation. too much to my mind that means having a robust discussion on the platform. people are going it is long been the case the best way to deal with lies is really ban them or suppress them but open them to sunlight and discussion social media platforms are a great place for people to disagree if you spent any time on them.

some have interesting things like exit.com community notes where you can have the robust discussion about the content of a tweets and push back on something you think is false. those types of tools, hope they continue to evolve i think the social media companies should look at how they can help out bottoman up discussion happen without trying to suppress the actual publication of content for. >> this is part one of your report for the other parts deal with? >> that will be updates over time. we are going to issue them 90, 60, 30 will do a postmortem after the election. we will do a postmortem after the election. then we will continue the same process we are iterating a little bit as the discussion around politics wraps up volume will as well.

week around a little. to identify all the situations in which ai were used in u.s. elections to build a database people can look back at and say i know about this incident i know what happened here is some evidence about what effect it had or did not have regrets working people find the forms? >> these reports are on our website. abundance.institute. you can go there and check it out. also learn a little bit more about the abundance. >> serves as a director of ai policy for the abundance. thank you for your time. >> coming up next on cspan2 from earlier today at infrastructure conference hosted by the u.s. chamber of commerce. then a higher education professionals from the university of maryland and the university of north carolina at chapel hill discuss the status of politics in america american university. later senator mitch mcconnell, dick durbin john cornyn talking

about u.s.