FTC Summit on AI 🗝️ Takeaways
On January 25th, the Federal Trade Commission held a Virtual Tech Summit on AI. The four-hour event was split between remarks from the FTC Chair, multiple Commissioners, two Bureau Directors, and panel discussions with industry representation.
The summit's central theme was to push the message that "there is no AI exemption to the law" and that companies will be held liable for the harm their algorithms cause to consumers. An undertone to all of the official remarks is the belief that FTC needs to act to prevent what it perceives as budding AI monopolies from forming. Fool us once…twice…not this time.
The Summit was also used as a platform to announce a compulsory 6(b) inquiry into the major AI platform companies. "The agency’s 6(b) inquiry will scrutinize corporate partnerships and investments with AI providers to build a better internal understanding of these relationships and their impact on the competitive landscape. The compulsory orders were sent to Google [Alphabet, Inc.], Amazon.com Inc., Anthropic PBC, Microsoft Corp., and OpenAI Inc." https://www.ftc.gov/news-events/news/press-releases/2024/01/ftc-launches-inquiry-generative-ai-investments-partnerships
FTC has not released a transcript of the event, so I used AI tools to summarize the FTC event on AI.
Below is a ChatGPT-enabled summary based on Google Doc speech-to-text transcripts. They are accurate to what the official remarks were.
If FTC does release transcripts, they will be posted here https://www.ftc.gov/news-events/news/speeches.
Highlights of FTC's AI Priorities
Chair Lina Khan
Char Khan's remarks at the summit focused on the challenges and opportunities presented by artificial intelligence (AI). She discussed the FTC's role in addressing these challenges, emphasizing the need to adapt legal tools and authorities. Khan highlighted the balance between embracing technological advances and protecting against their misuse, like fraud, discrimination, and surveillance. She drew parallels between historical and current market dynamics, stressing the importance of fair competition and warning against the concentration of power and control in the hands of a few dominant firms. The FTC's approach, she noted, includes focusing on business models, aligning liability with capability, and crafting effective remedies. Khan underlined that there's no exemption from existing laws for AI and highlighted the FTC's commitment to enforcing these laws to ensure a balanced development of AI technologies.
Commissioner Rebecca Slaughter
Her remarks focused the importance of foresight and proactive measures in regulating AI, to foster a competitive and safe market environment.
Commissioner Alvaro Bedoya
Remarks focused on the ethical use of AI, corporate responsibility, and fostering an equitable technology ecosystem.
Bureau of Competition Director Henry Liu
Remarks focused on the Bureau's commitment to regulating AI markets to foster innovation, protect competition, and ensure fair access for businesses of all sizes.
Recommended by LinkedIn
Bureau of Consumer Protection Director Samuel Levine
Remarks focused on the FTC regulating AI to protect consumers and market fairness in light of perceived harm from prior lax enforcement of new technologies.
FTC's Top 5 Priorities
Google Docs | Dirty Transcript
Hello everyone and welcome to the Federal Trade commissions Tech Summit on AI I'm Stephanie Nguyen Chief technologist of the office of Technology a year ago chair con and the commissioner is voted to establish the office of technology to ensure that the FTC remains Nimble and keeps The ground running to execute our mandate against a quickly evolving Tech landscape and since then we've hired some of the most talented technologists in the country committed to amplifying the benefits of tech curbing its harms and enforcing a law this moment comes at a critical time we're stacked against deep pocketed corporate incumbent who can hire the sharpest software Engineers designers and researchers in the world to boost and protect their bottom line and at the same time perhaps more quietly over the years there's been a unique shift in the application of what a technologist is in government Beyond building designing and deploying digital products and services at agencies like us digital service GSA and 18f law enforcement agencies like the FTC Consumer Financial Protection Bureau and the Department of Justice are firing up a surge of tech capacity on cases investigations policies and rules and as public servants we have our own bottom line that drives our work to Serve and Protect the American people to date we've onboarded a dozen brilliant technologists with skill sets to cover a broad swap of the economy including a software engineer in journalist who built an open source tour browser for iPhone a geneticist with deep machine learning expertise in health a privacy engineer Who develops safeguards for AI for kids and teens and an investigative data journalist who covered algorithmic discrimination as a gig economy Scrappy Relentless and adaptable the small But Mighty team is making the most of the resources we have We're working directly on the cases and investigations and engaging in the policy and Horizon scanning research and we're receiving input from the communities who have been overwhelmed by security breaches surveilled by hidden data Brokers and squeezed out by the companies who have outsized power today the FTC is convening the summit to examine layers of the AI Tech stack to understand the extent of competition across the various layers and sub layers and their potential impacts on consumers we will look at three layers hardware and infrastructure data and models and consumer applications the first panel will focus on the computational infrastructure of AI including computer microprocessors and cloud services the discussion will look into have dominant firms can have control over key hardware inputs like semiconductor chips and we've highlighted that cloud computing is a central part of the economy a request for information flag submissions that raise concerns about widespread Reliance on a number of cloud providers potential security risks and business practices affecting competition the second panel will discuss data and it's used in training AMR models we'll discuss the methods of data collection for model development and training and how they can have implications for competition and consumer protection your rigorous enforcement the FTC has highlighted how companies can collect use and abuse user data which can be magnified with the speed and scale of generative AI we've seen that third-party tracking pixels enable platforms to a mass analyze and infer information about user activity and we know that highly private data through voice recordings and videos can be used to train algorithms AI facial recognition technology can recklessly be used for surveillance purposes and finally the third panel would examine another area of long-time interest to the FTC including the impacts of how AI is being deployed in consumer applications we use many tools at the FTC to uncover these benefits and harms for digital artists sci-fi writers and musicians we recently held around table and published a report to amplify their perspectives on how generative AIS impacting their creative fields and to prevent Monitor and evaluate fraud and scans through ai-enabled voice cloning we're partnering with the Bureau of Consumer Protection on an exploratory challenge above all today we're eager to be learning we've got a full lineup today with speakers from diverse sectors startup Founders Civil Society Advocates researchers and government officials thanks again for joining and without further delay alternative
Chair Lina Khan
A fantastic Chief technologist Stephanie win for spearheading today Summit as Stephanie noted the office of technology is relatively new but I've just been so struck by how they've hit the ground running and really ensure that the FTC is on the front lines of tackling some of these new challenges that artificial intelligence tools are presenting from The Voice cloning challenge that they set up to convening our conversations with creators to make sure that we're continuing to learn and able to adapt and update the application of our tools and legal authorities as needed and then partnership with our bureaus of competition and consumer protection our team is really been working with great agility as we navigate this fast moving moment of technological opportunity and risks over the last 18 months the rapid deployment of artificial intelligence tools have captured the world's attention spurring some combination of all Wonder apprehension and fear we hear how these automated Technologies could open up the door to break throughs across Fields ranging from science to education making life better for millions of people but we've also already seen how these tools can turbocharge fraud automate discrimination and entrench surveillance putting people In Harm's Way more fundamentally we Face basic questions of power and governance will this be a moment of opening up markets to fair and free competition unleashing the full potential of emerging Technologies or will a handful of dominant firms concentrate control over these key tools locking us into a future of their choosing which of these potential trajectories AI will take is not in inevitability the outcome will be a direct result of policy choices that we make now zooming out for a moment virtually every large firm going back to us steel and the early 1900s to Alcoa in the 1930s to IBM and AT&T in the 1970s to Boeing in the 1990s and to dominant technology platforms today have argued that their Market power is good for America and that government officials have an offered agreed in the 1990s officials even reportedly threatened the Europeans with sanctions if they would not allow the merger of Boeing and McDonald Douglas as one White House advisor put it Aerospace was the only sector where we have a de facto national champion and you can be out and out on good for it and yet when you concentrate production as going for example has done you also concentrate risk and so today we see first hand some of the real implications of that to quote United Airlines CEO Scott Kirby that 1997 Mega deal is what led directly to the transformation of Boeing from a highly profitable world class engineering Enterprise to an ossified money losing Corporation with dangerous quality issues that we're now seeing first hand Boeing's Journey unfortunately is no different from that of many large corporations the policy makers have historically insulated from competitive challenges and whose Market power Mass the decline of internal capacity the difference between Boeing and many of these companies is that they're simply no masking airplanes falling apart in the sky and so I think this recent experience just really underscores the stakes of the decision we face as we choose between a future and we're continuing to consolidate control and power versus really enabling and empowering open markets in for a competition similarly we Face similar questions prompted by new technologies in the mid-2000s and Beyond set of the Web 2.0 era and unfortunately we saw that what began as revolutionary set of Technologies ended up concentrating enormous private power over what have become near essential services and through aggressive strategies to acquire or lock out companies that threaten their position a handful of firms solidified their dominance while locking in business models that we now realize came at the expense over a privacy and security lawsuits around the country have surface the heavy cost from the decimation of independent journalism to Serious harm to kids Mental Health today policy makers across government recognize the importance of learning from these missteps as we navigate the challenges and opportunities posed by AI at the FTC the rapid development and deployment of AI is informing our work across the agency as we look to promote Fair competition and protect Americans from unfair or deceptive tactics there's no AI exemption from the laws on the books and we're looking closely at the ways that companies may be using their power to throw it Fair competition or trick the public as part of this effort The Investments and Partnerships being formed between AI developers and major cloud service providers through using the agencies six be Authority where scrutinizing whether these ties enable dominant firms to exert undue influence or gain privileged access in ways that could undermine for a competition across layers of the AI stack as we continue this work a few key principles are top of mind first we are squarely focused on how business models Drive incentives just as we've seen behavioral advertising fueled the endless collection of user data model training is emerging as another feature that could further incentivize surveillance the ftc's work has made clear that these business incentives cannot justify violations of the law the drive to refine your algorithm to not come at the expense of people's privacy or security and privileged access to customers data cannot be used undermine competition we similarly recognize the ways that consumer protection and competition enforcement are deeply connected with privacy violations fueling Market power and Market power in turn enabling firms to violate consumer protection laws and our remedies will continue to focus on deleting the models themselves in addition to unlock will be collected data second we are squarely focused on aligning liability with capability and control this requires looking upstream and across layers of the AI stack to pinpoint which actor is driving or enabling the lottery game and is best position to put a stop to it what AI liability regimes will ultimately look like is still an open question but our enforcement experience in other domains will directly inform how the FTC approaches this work for example our recent robocall enforcement sleep not only targeted telemarketers and the companies that hired them but also looked Upstream to the lead generators and voice over Internet Protocol providers that enabled the illegal part telemarketing and in our recent work to combat scams we are holding Upstream payment actors like Walmart accountable for knowingly facilitating the fraud third we're focused on Crafting effective remedies that address the underlying business incentives and also establish bright line rules on the development use and management of AI inputs the FTC is making clear that some data is simply off the table for training models for example our recent order against Rite Aid bans the company from using facial recognition tools after it's Reckless application of the technology led to innocent people being falsely accused of shoplifting and our recent cases against data Brokers include bands on their using or monetizing people highly sensitive location data as we continue to establish rules of the road for AI it's also essential that we set clear boundaries on the content that can and cannot be used for scraping and model training the commission recently held the public workshop with creative professionals to better understand the types of guardrails that would help protect against creators work being appropriated and devalued by generative AI models including in ways that may under my favorite competition Our subsequent report and the create on the creative economy lays out our core concerns and how our authorities May apply in this space across our work we are making clear that firms cannot use claims of innovation as cover for law breaking and we've already made that clear through a number of blog posts and other signals that we're sending to the market to make sure the model is a service companies in Cloud providers and others throughout the AI staff know that there's no AI exemption from the loss on the books learning from our experience in the mid-2000s we're using the full scope of our authorities to make sure that these hard-learned lessons don't repeat themselves much is uncertain about what the precise future of this technology will look like but the good news is that we have the experience and expertise to meet the moment and by continuing to sharpen our thinking and Faithfully enforce the law we can unleash AIS potential benefits from the potential Harms I'm so grateful for today's convenience the types of insights they'll be shared today will directly inform how the FTC approaches are work and so just want to give a big thank you both to our teams and that help put together today's event and all the speakers who took out the time to come speak with us and with that I'll turn it over to our deputy chief technologist Alexander will be leading our first panel.
Commissioner Rebecca Slaughter
Really grateful for this opportunity to speak before the incredibly distinguished group of panelists coming on the AI and data models panel and Stephanie's team has said where to pivotal moment the emergence and Rapid deployment of AI technology and other Advanced algorithms there are stories nearly every day about their potential to transform Industries and even change the work of government almost every facet of these developments has major implications for our mission to protect consumers and promote competition but we've also seen this play before we as a society as well as our specific agency are still dealing with the Fallout from a government-wide largely hands-off Approach at the beginning of the big data and social media area by learning from the past we can unlock the promise of this new technology make sure it is free from the control of just a few Gatekeepers to ensure it's safe deployment and use for all I see the story of our regulatory posture to the social media at Tech and Commercial surveillance era as in holding in three parts act one was the emergence of social media social media powered by Big Data and it was filled with nearly boundless optimism about our new foundability to connect and conduct Commerce easily and quickly over despite early warnings about privacy and Market consolidation by a few advocates regulators and legislators please only the most egregious conduct Act 2 in the 2010s we saw users and whistleblowers and experts bringing alarm Bells about data privacy arms to kids and teens and about the corrosive effect these large firms were having on our politics and markets still for most of the decade Regulators were playing catch up now Racine the consequences of the hands off approach the markets of Consolidated into and around a few extraordinarily large companies as these companies have gotten larger and control the advertising and attention markets the ones vibrant and iconic American journalism and arts sectors are in crisis the business model of many social media firms profits off our most private information And at the same time the content that does proliferate in these networks has facilitated a Teen Mental Health crisis and widespread misinformation and disinformation online that threatens the Integrity of Elections and our political system today we're acting in Earnest and with urgency to address these issues but it really isn't open question whether a vision of our digital ecosystem is a dynamic and open space to connect and share information across boundaries can be restored after a problem is entrenched is often way too late to correct it we're in the middle of this play when it comes to the data-driven AI and algorithmic Technology markets this is a moment of both promise and peril we've seen the deployment of advanced algorithms to make decisions in healthcare criminal justice Employment Credit housing and other areas of economic consequence through vigorous strategic enforcement of our competition and consumer protection laws we can begin to meet the challenge of this new era and create a Marketplace that unleashes Innovation while protecting consumers and competition and AI Marketplace competition can be hampered by limited access to key inputs such as Processing power and chips as we already in the previous panel and training data as I imagine will here in the next these key inputs can be controlled by large incumbents I worry that the concentration of AI models with access to huge amounts of consumer data and a hands of very few companies compose enormous risk including around consequential economic decisions and access to opportunity information and privacy if even a little of the hype around the power of AI models is to be believed than it is emphatically our responsibility and obligation to support open and fair markets and prevent monopolies in their incipitency we should also all take note of the proliferation of Partnerships and direct Investments involving AI developers and large technology companies that are structured to avoid triggering notification to the FTC or doj and he trust division under the pre-merger notification rules it is reasonable to wonder whether these Investments could lead to a heavily Consolidated market dominated by only a few cars with either no competitors or with competitors who are hamstrung by their dependents on those same incumbents that's why I was very pleased to support the 6B study that the chair announced that the start of the summit the way to stay on top of this quick moving market and avoid repeating the mistakes of the past is by using the full piano authority to make sure we are fully aware of the business models and related incentives and consequences that are taking shape in the world of AR of course studying a market is a compliment to and not a substitute for appropriate enforcement if the laws and the facts supported there is no AI exception to the law even Investments that do not have to be noticed under the hard spot where do you know act procedures May violate the underlying antitrust laws I am confident that where there are facts that support enforcement investigations our agency will pursue them proactive use of our Section 5 consumer protection authority will also be essential to unlocking the benefits of a competitive AI Market while avoiding the mistakes of the past I'm a Believer in the potential for AI tools to help make our lives easier but as we saw on the commercial surveillance era pipe Cycles Drive Investments and products and deeply inform consumer purchases when that hype is build on a foundation of unsubstantiated over promises about a product's capabilities billions can be lost honest marketing claims are deeply Pro competitive deceptive marketing about an AI products capabilities crowds out scrupulous firms that promise only what they can actually deliver I also continue to be concerned about the intersection of data privacy and access to economic and Commercial opportunities and bias decisions those can stem from differential and accuracy for different demographic groups as we've seen in biased facial recognition technology from the bias deployment of those tools and marginalized communities to the repetition of existing patterns of economic distribution under the guise of unbiased algorithmic decision making I'd like to see us write a different play than the one we saw involved in the commercial surveillance era nothing about her basic data collection eye tracking the shape of social media or the dominance of a few Tech firms was inevitable in action in the face of those developments was a policy choice we have the knowledge and experience now to see this era play out differently I'm so excited to learn from the leader this year today about how we can build a vibrant safe and competitive market in the AI era I'd like to hear about the lessons we can learn from that first digital Revolution thanks again to the team of the FTC and to our guest for being here.
Commissioner Alvaro Bedoya
I'd like to turn it over to commissioner bedoya Now for Something remarks commissioner I want to thank you Stephanie everyone at the office of Technology who has spearhead of putting this event together I want to thank all the staff I would learn from and who I've had the pleasure to work with so I'm really really glad to be here I'm also glad to have voted out the 6p order to shut some light on the competitive Dynamics to play with somebody's most advanced models I will probably Focus Less on some of the generative and LM models that have been the focus of the discussion today but I do want to touch on One Note I think is just so critical for a regular people today Jonathan Frankel who is an old friend and colleague mentioned this trend that he seen in some of these llm models in generative models which is a trend toward less transparency around the data that is being used and you look from a few years ago there's actually quite a degree of specificity around what data is being used and nowadays A lot of them have have pretty noticeably shifted to little to no transparency around what data is being used and I understand there's a variety of reasons for that but I'll just say that from the standpoint of bias of detecting Trying to suss out whether these systems work differently for some groups of people rather than others this has consequences because knowing what data is used to train the train system up is extremely valuable for understanding how that system might from why do you just hit myself down In the future with respect to bias and so I want to focus on that issue on the issue of bias in AI systems and as I as I previewed I want to focus a little less on the generative and llm models that have been discussed not cuz I'm not interested I obviously I am but partly because they've been the subject of discussion but most importantly because I think that people today are much more likely to encounter in their daily lives a much more traditional automated algorithmic decision-making system then they will a generative or llm system and they are exponentially more likely to have critical decisions made about them by those older or traditional algorithmic automated decision-making systems and so today I'd like to focus my attention and my remarks on just a bias in in those systems and focus on just one case which is our recent Rite Aid settlement which came out I think of December 23rd of last year terrific work by our division of privacy and identity protection and our Bureau of Consumer Protection more broadly and I'll urge you to look up the complaint look up the settlement get the details but in a nutshell this was a case of the retail pharmacy using a face recognition system to try to identify suspect the chocolate lifters who were entering its stores unfortunately as we allege in our complaint the system was flawed it was we alleged profoundly flawed and despite all of this it was we alleged disproportionately deployed in areas that were plurality minority in which the system was we alleged most likely to misfire and the ways in which it misspell hired were were pretty jar and I'll just share a few instances here with you today one instance a right-aid employee Stockton searched at 11 year old girl because of a false match this girl was so distraught that her mother said she must work in order to comfort her this 11 year old child who was falsely stopped in a in a store in another instance Right Aid employees call the police on a customer because the technology generated an alert against the image the customer they stopped was African American in countless other instances people were stopped shopping for food shopping for medicine and other Basics they were wrongly searched they were accused of shoplifting and in many instances expelled from stores and sometimes all this happened in front of their families in front of their co-workers in front of their bosses and so why am I raising this all of this now what what takeaways do I want people who are watching businesses and their counsel who are watching the take away three things That Corey mentioned in the last panel which is we need to appreciate what's at stake here and he had this great line that you don't call a murder effect and I'm not using anyone to murder here but I do think that this is a reminder that that yes you know this is about privacy you know yes this is about right here but it's also about people walking into a pharmacy to get their prescriptions and being stopped have having the cops called on them and being asked to leave because a secret face recognition system has wrongly identified them as someone they are not the decisions made by the systems and and zooming out more broadly from Rite Aid affect our basic ability to live our lives with dignity with fairness to get the healthcare we need to get fairness in our in terms of the apartments we rent the jobs we apply for in countless other areas of Our Lives this technology has basic implications for our lives in our abilities to lead them that's the first thing the second thing I want companies to take away is that a company cannot buy a technology a company cannot deploy it on on its customers and masks and when that technology cannot turn around and say I am sorry it was an algorithm it was a black box I didn't know I am sorry I am not liable if you are a technology that is using an algorithmic decision-making system and this is the third takeaway to make critical decisions about people's lives in a way that may substantially injure them you need to ask hard questions about how that system works how it affects people how it can hurt them and you need to make sure that system works fairly and is fair to the people it is used against I'm in in closing I I want to to actually answer if I may be the closing question that krisha asked our moderator from the bureau competition from Chad asked to close last the last panel too because it really helpful question which is what does look like and I'm going to get you answers to this the first you know I think success looks like people controlling technology not the other way around success like it looks like people feeling like they're in control of Technology people knowing when technology is being used to make decisions about them people knowing why those decisions were made people knowing what they can do if those decisions misfire and people generally appreciating what is happening around them and why technology not the other way around and the second answer I would give here is more from the competition language I know is rightfully the focus of most of our panels today and and I think that from a competition side success looks like us using a technology not because the biggest or most powerful company put it out and put it right in front of us using the platforms we already use but rather us using a technology because it is proven itself to be the best technology in the marketplace on its merits a success looks like people like Jonathan or other startups being having a shot at success having a shot at reaching customers because they put out good products that work that people like and not just because those products are put out by the most powerful company that has hit a trillion or more dollars in valuation success looks like companies on the merits not because they are large and so with that let me turn it over to Andy Hasty from our division and privacy and identity protection and I just am excited to watch our next panel which.
Bureau of Competition Director Henry Liu
Segment we will hear from our Bureau of Directors for Stuff I'd like to introduce our Bureau of competition director Henry Lou for some closing statements Henry over to you good afternoon everyone what a fantastic Summit thank you Stephanie Wen in the entire office of technology for organizing this event it's been really great to engage with such a diverse group of stakeholders to discuss Ai and the many layers of technology related to it got a hearing panelists cover topics such as the potential for dominant firms to gain control over key inputs and use data without consent it's a reminder of the increasing significance of AI tools as we close out the summit I'd like to share my views on the Bureau of competitions role in this conversation possibilities to identifying National Security threats but there are potential risks too often AI tools have been used to limit opportunities and prevent access to critical resources to guard against potentially legal Behavior while also allowing the benefits of AI to flourish it's extremely important for the antitrust agencies to be vigilant to protect competition in the Bureau of competition our role in this new AI power rule is clear we will please illegal behavior and vigorously in the cost of an action can be outsized in AI markets because excessive Market power can distort the path of innovation and discourage future investment in AI research and development so what principles will guide the Bureau of competitions enforcement priorities for AI markets I'll mention a few competition policy have long help treat conditions for fair and honest competition as we heard from several panelists this afternoon when dominant firms control over key inputs like computing power cloud storage semiconductors talent and even data it creates a risk of excessive crisis and course of terms to promote more open markets the bureau will continue litigating cases that address conduct and acquisition switch if left unchecked might result in increase months Market concentration in various layers of the AI staff to give you an example in June 2021 the FTC Suite broadcom alleging that it illegally monopolized markets for Semiconductor components or chips used to deliver television and broadband internet services then in December 2021 the FTC sued to block new videos proposed 40 billion dollar acquisition of Chip design provider arm protecting competition Hardware will help Propel AI develop most recently the commission empowered Bureau staff to ramp up enforcement by authorizing an Omnibus resolution which allows us to use compulsory process in non-public investigations involving products and services that use reclined to be produced using AI the Omnibus resolution will streamline our ability to issue civil investigative demands which are a form of compulsory process similar to a subpoena an investigations relating to AI the second principle and along with protecting fair and open markets is that AI tools should support opportunities for new businesses to compete according to a survey by the Small Business and Entrepreneurship Council many businesses are already using AI tools to improve their efficiency and save billions annually from marketing and sales to Human Resources AI tools of course have a wide range of capabilities with such broad applications there of course often significant downsides attached for example small businesses that want to use AI tools frequently find it cost prohibitive in part because they also need to acquire a hardware software and skilled professionals ensuring that markets for these critical inputs remain competitive and support a diverse ecosystems of startups working on AI projects now keeping markets competitive for all businesses will require that the bureau continued using all the tools in its toolbox to vigorously enforce the antitrust laws one important tool is section 5's unfair methods of competition Authority in November 2022 the commission issued a policy statement on the scope of Section 5 Authority and in that statement the commission made it clear that it is congressionally mandated to identify unfair forms of competition including those were small businesses were threatened by oppressive Powerful rivals more broadly the ftc's unfair method of competition Authority empowers us to protect the public against powerful firms that unfairly use AI Technologies in a manner that tends to farm competitive conditions in AI or AI adjacent markets the commission must remain Vigilant to protect the public against such threats and to ensure that competition thrives in AI input markets as well as in markets impacted by new and emerging AI applications with these enforcement Activities The Bureau and the commission are doing what they have done for over a century that is to keep Pace with new markets and Ever Changing Technologies are vigilance has been in no smart small part due to our office of technology and their in-house expertise on all things Ai and I look forward to our continued joint work with that I'll pass it back to
Bureau of Consumer Protection Director Samuel Levine
[Copied from official speech posted to FTC site]
Thank you again for joining the FTC’s first Technology Summit on artificial intelligence.0F 1 I’ve been so impressed by the thoughtful, in-depth conversations we’ve had around topics ranging from cloud infrastructure to privacy to AI-fueled fraud. And I’m so grateful to our Office of Technology for organizing this important event. Although formed less than a year ago, OT is already having a huge impact on the day-to-day work of our agency. In the bureau I lead, the Bureau of Consumer Protection, OT technologists are partnering with us on dozens of matters across many offices and divisions, and they are helping to ensure that our enforcement and policy work can meet this moment. And let me be clear – this moment is unique. Chair Khan spoke earlier about the last major inflection point driven by emerging technology, at the dawn of Web 2.0 in the early 2000s. She noted how that era began with promise, with exciting new technologies and applications connecting people and expanding opportunities. But two decades on, we see a tech ecosystem that has concentrated private power in the hands of a small number of firms, while entrenching a business model built on constant surveillance of consumers. Chair Khan stressed that we need to learn from our experience in that era, and I could not agree more. This learning is not an academic exercise. As we chart our course in confronting AIrelated harms, we need to engage with the history of how we arrived at this moment. A generation ago, in 2000, the FTC issued a major report finding that self-regulation on the then-nascent internet was failing to protect consumers’ privacy, and the Commission recommended that Congress pass legislation.1F 2 Yet less than 18 months later, with a change in administration, the FTC reversed itself – finding “it is clear that industry will continue to make privacy a priority[,]” and warning that legislation would be premature and could hold back the growth of the internet.2F 3 Instead, the FTC announced an initiative to ensure privacy policies were posted and honored, and to encourage industry self-regulation.
In my view, this reversal was a serious error. It is now apparent that industry did not actually make privacy a priority. In fact, at the same time the FTC was expressing confidence in self-regulation, Google began exploring how it could mine search queries for behavioral insights on its users3F 4 – laying the groundwork for a transformed business model, and ultimately a 5 transformed internet.4F A decade later, the FTC would reverse itself once more– calling in 2012 for Congress to pass legislation5F 6 – but by that time many of the harms Chair Khan described were already entrenched, and industry opposition to meaningful regulation had consolidated. This history is very much top of mind as we confront the emerging threats and opportunities created by AI. Congress has entrusted us with ensuring that markets are fair and competitive, and we will not be sitting on the sidelines. In the Bureau of Consumer Protection, we’ve been staying busy. We issued a major report in 2022 – months before generative AI became a hot topic – warning about inaccuracy, bias, and privacy abuses fueled by AI.6F 7 In 2023 we began routinely issuing guidance on how our authorities apply to emerging AI technology, making clear that we were not going to wait and watch while harms accumulate.7F 8 And as we’ve talked the talk with our guidance, we are walking
the walk with our policy and enforcement strategy. We’ve now required across multiple cases that models trained on illegally collected data be deleted.8F 9 We’ve brought lawsuits against firms that defraud the public by claiming AI can make people rich.9F 10 We’ve partnered with the Office of Technology to launch a voice cloning challenge to confront new forms of impersonation fraud,10F 11 and we have proposed a rule to prohibit and deter the practice.11F 12 We’ve made clear that firms can’t retain kids’ data forever, even and especially to train models.12F 13 And we’ve established that firms must either take steps to ensure their AI tools don’t harm consumers, including by discriminating against them, or cease to use these tools altogether.13F 14 None of this is to suggest that more resources and authority are not needed. But what should be clear is that we are using every tool – enforcement, rulemaking, education, market studies and more – to protect the public from emerging harms. Our most important tool is our people – our multidisciplinary teams of world-class attorneys, economists, investigators, consumer education specialists, and technologists – and all of us benefit enormously from events like these that engage top experts from both inside and outside the government to better understand how AI is reshaping the marketplace. A generation from now, when a future Bureau Director – or their AI avatar – discusses the history of this era, I am confident they will recount an FTC that was active and engaged in ensuring that AI’s promise can be harnessed for the benefit of people, rather than a handful of tech giants.
AI and Tech Legal Counsel | U.S. Licensed Attorney | GenAI Prompt Engineer
1yThe FTC is acting on AI - “The agency’s 6(b) inquiry will scrutinize corporate partnerships and investments with AI providers to build a better internal understanding of these relationships and their impact on the competitive landscape. The compulsory orders were sent to Google [Alphabet, Inc.], Amazon.com Inc., Anthropic PBC, Microsoft Corp., and OpenAI Inc." https://www.ftc.gov/news-events/news/press-releases/2024/01/ftc-launches-inquiry-generative-ai-investments-partnerships