AI Deception Machine: Deepfakes, Chatbots, and New Frontier of Senior Fraud

Special committee on aging will now come to order. Thank you all for being here today. One of the biggest issues I hear about from Floridians and seniors around the country is the growing threat of scams, fraud and financial exploitation. Whether it's a phone call from someone posing as a grandchild in trouble, a suspicious investment scheme delivered through the mail, or an email from a government imposter threatening jail time. These criminals are targeting their seniors with increasing sophistication. Our seniors are often especially vulnerable to this kind of fraud. Sadly for many older Americans, falling victim to a scam doesn't just mean losing money. It could also mean losing peace of mind, trust in others, and confidence in themselves while navigating daily life. As this committee has heard many times bef- before, this is a multi-billion dollar a year problem. Twenty twenty-five Americans over sixty lost a staggering seven point seven billion dollars to scams. And that's just what scams they've been reported. The rise of artificial intelligence has offered scammers new tools to pursue their criminal schemes, and we must adapt and respond to these new threats. AI can be used to make fraudulent videos known as steed fakes, impersonating experts and celebrities to try to fraud our seniors in a variety of devious schemes. AI can also be used to clone someone's voice, a terrifying development that has been list- used in heartbreaking and evil ways to impersonate a loved one and deceive their family. AI has been used, has made existing scams more effective and lowered the barrier to entry for criminals. But once required, real coding skill cannot be done by anyone with an AI tool. Many of these actions are highlighted within the fraud section of our new report, Artificial Intelligence in Older Americans, Confronting New Threats, Unlocking New Opportunities. I am grateful to the ranking members of the brand for working with me to put this report together. The report, which is available online at aging dot senate dot gov, includes helpful information educating on the new threats of AI and efforts you can take to protect yourself. Many seniors live on fixed incomes, and this kind of exploitation can be the difference between a secure, comfortable retirement and years of financial hardship, distress, and isolation. That is why we are focused on combating fraud at every level. We must empower seniors, families, and communities to protect themselves, and then we must act to fight back. That means understanding this new technology, supporting our law enforcement, educating the public, and getting innovative about how we protect vulnerable Americans. Understanding has to come first. That is why I introduced the bipartisan Aging with Artificial Intelligence Act of twenty twenty six, with Senators Mark Kelly and Roger Marshall. This bill directs the national academies to study how older Americans are using AI and to weigh both its benefits and its risk, including the various scams and fraud we are examining today. Before we can protect seniors from this tech we need to understand how he's reaching them. I'm gr- grateful this effort has the support of groups like AARP, the American Medical Association, and the American Psychological Association. But only understanding the threat is not enough. Seniors need an easy way to report it. That's why Senator Maggie Hassan and I introduced the Report Scams dot Gov Act, a bipartisan bill that will establish a single federal portal where consumers can report scams find information about available assistance and have the report automatically routed to the relevant federal and state agencies. Seniors should not have to navigate a maze of bureaucracy just to be heard. Of course, a report is only useful if law enforcement can act on it. That's why I also joined Senators Katie Britt and Ranking Member Jill Brand to introduce the Guarding Guarding, Unprotected Aging, Retirees from Recep Deception Act, known as the Guard Act, to expand existing federal grant programs. The bill funds specialized law enforcement training on elder fraud and provides tools to trace stolen cryptocurrency. Finally, these efforts must be coordinated. That's why I introduced the National Strategy for Combating Scams Act with ranking members of the brand to require the FBI to develop a unified national strategy informed by victims, law enforcement, nonprofits, and the private sector. The federal government has to coordinate its efforts to effectively combat these complex schemes. Each of these bills attacks the problem from a different angle, but they all share one goal, making sure seniors are never left to face these criminals alone. Today's hearing is part of that effort. We will hear from experts on AI policy, leaders from the financial sector at work to stop these thugs, and victims who have faced this, these crimes firsthand. Our goal is clear. We want to educate seniors about the rising threat of AI-enabled fraud, show them how to defend themselves, and identify the policies the federal government can enact to help curb this threat. Seniors deserve to feel safe when answering the phone and opening the mail. They deserve to know that when they report fraud, they will be heard and action will be taken. Protecting them isn't a a partisan issue, and I'm grateful for the partnership I have with the members of this committee. I'd now like to recognize recognize ranking member, Jill Abram, for her opening statement.

Thank you, Mister Chairman, and I appreciate all the witnesses being here for such an important hearing. The question before us is not whether AI will transform our society, it certainly will. The question is whether we let it be turned against the people we love. Right now, criminals are using AI to steal from families and seniors. cloning voices, faking videos, and tricking honest people. Americans deserve to trust their own eyes and ears. That takes clear rules, honest labels, and real consequences. In just a few years, AI has become part of daily life for nearly every American, older adults included. AARP found that AI use among older adults nearly doubled between two thousand and twenty-four and two thousand, twenty-five. That's no surprise. AI holds real promise to improve seniors' lives. In New York, the Office for the Aging is already connecting seniors with AI tools that fight scams, ease loneliness, and support caregivers. We should encourage those uses. But we must also put common sense safeguards in place so this technology is never used against the people it is intended to serve. I often think about Pope Leo's first encyclical. Magnifica Humanitas. His call to safeguard human dignity in the age of AI. Pope Leo warns that technology is never neutral. The same tools that connect us can also be turned against the most vulnerable among us. Today, we will hear how that is already happening. Scammers use AI to generate deep fake images, clone voices, and launch phishing attacks. that drain seniors' savings. In many cases, AI didn't invent these scams, it supercharged them, making old cons more convincing and far easier to pull off. AI deception takes other forms too. Tools, the tools that are built without real safeguards can hallucinate and feed older adults dangerously wrong information. Last year, Reuters reported an AI chatbot that repeatedly insisted it was human and invited a seventy-six year old man to an address in New York City. He fell on his way to the train station and died of his injuries. This year, the New York Times reported that an AI tool gave a seventy-five year old man bad medical advice, and he delayed his cancer treatment because of it. Even one case of AI fuel deception is too many. We can choose differently. We can decide how AI shapes our communities, our families, and the way we care for one another. This morning I sent a letter to the leading AI companies, to the AI companion companies, including OpenAI, Meta, and X, asking them to explain how they're sa- they safeguard their products for older adults and people with disabilities. I was glad to partner with Senator Kelly on this very letter, and I expect these companies to respond promptly and fully. That letter is just one step. It can't be the last. We need a bipartisan framework to govern AI, one that heeds Pope Leo's counsel by putting human dignity first and protecting the most vulnerable including older adults. We must also confront the scams targeting our seniors head-on. I've introduced two bipartisan bills to do exactly that. the National Strategy for Combating Scams Act and the Guard Act. Both have federal agencies and law enforcement, both give federal agencies and law enforcement the tools they need to coordinate and to protect consumers from fraud. I look forward to working with Chairman Scott and my colleagues on this committee to ensure that our older adults are protected from AI's harms and share in their benefits.

Thank you, ranking member. I'd like to welcome our witnesses today. They are here to help the committee understand the landscape of AI driven fraud, and share their experiences dealing with these types of scams. First, I'd like to recognize Doctor David Amron, an internationally recognized surgeon and the founder and medical director of the Roxbury Institute. He joins us today because he saw firsthand how this technology can be turned against us. Scammers used AI to create a deepfake video of him endorsing a fake uh miracle cream a product he never made and never endorsed thank you for being here you may begin your testimony

thank you chairman scott chairman scott ranking member jillibrand and distinguished members of the senate special committee on aging thank you for the opportunity to testify before you today my name is doctor david amron and i'm the founder and medical director of the roxford institute in los angeles california and the founder and chair of the lipoedema society For more than three decades, I have dedicated my career to treating patients with Lipidema, a chronic and often misunderstood disease that affects millions of women worldwide. I am widely recognized as a pioneer of Lipidema surgery in North America, and have devoted my career to advancing treatment, educating patients, and advocating on behalf of patients to help them obtain insurance coverage for their care. As physicians, trust is the foundation of everything we do. Patients rely on us during some of the most vulnerable moments of their lives. I have spent more than three decades building that trust, yet scammers used artificial intelligence to exploit it in a matter of moments. Although I know AI was rapidly evolving, I never imagined how quickly it could be weaponized to deceive patients, until it happened to me. We first became aware of the scam in the summer of two thousand twenty five, with reports increasing significantly by early September. Patients began contact contacting the Roxbury Institute after seeing a video that appeared to show me endorsing a so-called " Miracle Cream". "Miracle Lipidema Cream". Many forwarded us the video, but by then, some had already purchased the product before realizing it was fraudulent, including several of my own patients. The video was disturbingly realistic. Scammers used actual footage from my YouTube channel and combined it with fully integrated ai generated likenesses of my voice and my colleague dr. karen herbs the head of research and director of diagnostic and private medicine at the rock students too as well as ai generated celebrity images and stolen media logos to create a polished advertisement that appeared entirely legitimate as soon as we recognized it a coordinated defake as a coordinated defake my team immediately issued public advisories across social media platforms to warn patients clarify that neither doctor herbs nor i had any involvement and help our community recognize and report ai generated medical impersonations our web developer and digital marketing strategist immediately began investigating the scam as soon as patients alerted us before taking action he investigated the company behind the product speltra minestra and quickly confirmed it was not a legitimate legitimate medical entity he reported every version of the advertisement to meta the fraudulent accounts responsible for publishing it, the domain registrars hosting websites, and the major search engines directing consumers to the scam. Yet each time one version was removed, another quickly appeared. This continued over the course of eleven days. Despite these persistent efforts, the scam kept on recurring. It became quickly clear that existing reporting scams, eh, reporting them was not simply enough to stop a coordinated AI-enabled fraud, campaign. After exhausting every avenue available to us, we realized we needed to bring a national tension to what was happening. We then reached out to major television networks, because we believed public awareness might accomplish what traditional reporting mechanisms could not. The Today Show responded, conducted its own investigation, and ultimately brought the scam to a national audience. The Today Show attempted to identify and cont- and contact the individuals behind Svelte Vinostra, just as we had. Like us, they found no legitimate company, no accountable representative, and no one willing to respond. Although the fraudulent video was eventually removed, versions of the scam have since resurfaced, underscoring how persistent and difficult these AI enabled fraud schemes are, to eliminate. What disturbed me most was not that somebody had misused my image, it was knowing that scammers had used my name and my reputation to deceive the very people I have dedicated my career to helping and protecting. The consequences extend far beyond financial loss. Patients may delay legitimate medical care for a progressive disease like lipedema, place their trust in unproven products and services, and allow their condition to worsen. Perhaps most damaging, these scams erode the pr- the trust patients place in their physicians. Once that trust is broken, patients no longer know whom they can trust. Artificial intel- intelligence has fundamentally changed the landscape of fraud, the ability to convincingly impersonate trusted physicians, fabricate endorsements, and deceive patients on an unprecedented unprecedented scale. What happened to me is not an isolated incidence. Physicians across country discovering their images, voices, and professional reputations are being used without their knowledge or consent to promote products they have never evaluated, recommended, or many times even heard of. The danger is clear. These scams put at risk and undermine the credibility of the medical profession. Older Americans are especially vulnerable because they often place tremendous trust in their physicians. Many also rely on the internet, including social media and online search results, to learn about medical conditions, make them especially susceptible when fraudulent content appears to come from trusted medical professionals. When th- when they see a realistic online video of a trusted physician recommending a treatment, they have little reason to question its a- authenticity. That is precisely what makes today's AI-generated deepfakes such a powerful tool for deception. This problem is no longer theoretical. It is happening right now, and the consequences for patient safety, medical ethics, and the integrity of healthcare information are profound. If we do not act more, older Americans and other vulnerable patients will become victims of increasingly sophisticated AI-enabled fraud. I respectfully urge Congress to strengthen protections against AI-generated impersonation scams, improve accountability for those who create and distribute them, and ensure our laws keep pace with technology that can so easily be used to exploit trust for financial gain. Thank you for the opportunity to testify today. I look forward to answering your questions.

Thank you very much. Sorry that happened to you. Um, next I'd like to introduce Deborah Del Mastro. Uh, she's a Bay Area mother from Martinez, California, who became a public face of the growing AI voice clothing scam after criminals used artificial intelligence to mimic her daughter's voice. She is also a San Francisco Bay Area singer, voice coach, actress and musician who's taught voice, drama and trumpet for over three decades, sharing her voice and her love of music with her community. Thank you for being here and please begin your testimony.

Thank you. I I am uh very honored to be able to come and share my story with you also. Um I think my main my main goal is to let people know that these criminals have have tech you know uh technology that's uh that we don't know about i called this the worst day of my life it was may fourteenth um uh thursday may fourteenth the day started with a phone call in the morning we were my husband and i were at the breakfast table um and i answered a phone call from a local number Uh, I get a lot of my work that way. I get, you know, and um, I didn't think anything of it. Um, I answered and said hello, and a male voice said, " Hello, who is this?" And I said, " Who is this?" And he said, "Someone you need to talk to. I have your daughter." "Does your do- is your daughter prone to panic attacks?" Again, I'm sitting in my pajamas at the breakfast table. And I said My daughter? Uh, yes." And then he put on, I could hear commotion in the background, he put on her voice, now I know her voice, and not her, um, saying, " I'm so sorry, so I'm so sorry, mom, I'm so scared, I'm so sorry," and crying, sobbing. Um, it was her voice. It was absolutely her voice. And then he got back on and said that, uh, he had ten pounds of cocaine in his truck. in his trunk, and my daughter saw a deal go down that she shouldn't have seen, and that his boss was Mexican drug cartel. And uh his boss told him to take her, and he says um his boss doesn't care who he kills, and that he wants twenty thousand dollars for the return of my daughter, or he will sell her for what's between her legs. And he will kill me, my family, And and my daughter, if I don't do this, if I don't come up with the money. Now, we don't have that money. We just don't. Mmm. So I'm I'm incredulous with my jaw dropped. Um, I couldn't speak to my husband. Um, I quickly got dressed. I have him on the phone the entire time. And as I left the house, I mouthed to my husband that she'd been kidnapped. Um, he didn't know where I was going. or what was going on at all. Um, but he could just tell it was serious. This man was barking orders to me about having to have the money right away or the boss will do something with her. He told me to bring a phone charger and not to let my phone die or else I would never see my daughter again. He made demands for the next five and a half hours. First, to send two thousand dollars by Western Union to Mexico. Um, he wanted twenty five hundred dollars. I told him I didn't have it, which was really true. Uh, just two thousand. He had me go to a local Walmart and, um, I, that's where I used MoneyGram to send two thousand dollars to Mexico. I took a picture of it and texted him the receipt as he asked. And, uh, then he told me there was a problem with it, and that there was an error and I needed to spend, send the money again. I told him I did not have the money. Um, and he said, uh, that, uh, I would be reimbursed, it would be fine. Just send it again. And I, I, I said I, I did not have it. He wanted another five hundred dollars. I had it. So, I said, alright, and I went to my car and found that in my panic, I had locked my keys in the car. Then I had to call my husband to come pick me up, which also brought his bank accounts into play. And, uh, my husband drove us around with his phone on gps and my phone on with this person listening to every every moment of it um he ran us around to a bunch of different places uh twenty twenty miles away to another walmart um and to martina's twice we were on the phone with this awful man the entire time listening for every sound asking where i was what was going on if I was silent, threatening my family's lives and telling me that he could sell my daughter right now. I tried our local Safeway to send it, it wouldn't work, um went to Walgreens, he told me not to speak during the entire time I was in any store doing these um these payments, and to leave the phone on so he could hear everything that happened. Again I took pictures of the receipts and texted him, now a different number, um all the time he was threatening if I didn't hurry. there was gonna be problems. Now he said he was gonna drop my daughter off at the first place, then he was gonna drop her in another place, then he was gonna drop her off in another place. Um as we were uh uh driving to different places and sending money, we sent uh we sent fifty four hundred dollars altogether in four different um four different transactions. Um when it was over, it was about five and a half hours. Uh by the time they probably figured that we didn't have any more money to give them. Um, he ended it. He said, " It's over." We were now back twenty miles away at another Walmart. And he said that he had, um, released my daughter at the front of the Walmart and we were there. Uh, and he said, " Go, go get her. Go pick her up or call her." And hung up. Uh, I jumped out of the car. She was not there. I was in a total panic. And then I called her. And she answered the phone and said, " Hi, mom. What's going on?" And I was never so happy in my life to be scammed, so relieved, and furious that we'd fallen for this. Um, I heard her twice. Um, I had to be uh silent with uh even in the car driving, there was one time I actually muted the phone so I could speak to my husband. And he started screaming not to do any fuzz, no funny business or I'd never see her again. Yeah, he threatened to harm us. And my daughter saying that he had kids and he would kill anyone that got in the way if his kids were at risk. And his boss didn't care about anyone. He talked many times how he would sell Sarah for twenty thousand dollars, or he could, she could pay her way in other ways, hinting that she would be sex trafficked. But he was doing us a favor because we were respectful of him. He had me talk with Sarah again. I asked many times if I could speak to my daughter to make sure she was okay. Um, he had me speak to her again after we had sent, I think, the third, third amount. And, uh, he told me exactly what I was to say to her. He said, " Tell her that this nightmare is almost over. You're gonna pick her up as soon as possible." I did that. I said exactly that, and my daughter's voice said, "I'm so sorry, mom. I love you." After that, another man's voice came on who was much more hard, much more harsh, the enforcer, who threatened harm to us and Sarah if we didn't come up with more money. And that's when we sent another twelve hundred dollars. It was the last of our available accessible money. Um, it was the worst day of my life, and my husband, too. Um Yeah, some people are just so broken. Um, I'm a Navy veteran. I am uh, I am usually very, very calm and collected. In the face of crisis, I always say I'm the person that runs to the fire, not away from it. And I was totally, totally convinced that this was my daughter's voice. Yeah, uh, you, you know, when y- It's the perfect scam. It's the perfect scam. to uh, talk to family and tell them that you're gonna kill them, sell them. You know, it's a perfect scam. Cuz you're willing to do anything to, to get your, get your, your kids, get your daughter back. I spent most of that time in silence with my husband driving back and forth, just trying to figure out how he's gonna help her get through her life after an event like this. And thank God, it was a scam. Thank God it was a scam. But, I had no idea that this was a possibility, that, that AI could clone the voice. Now, when I called my daughter, both my kids work in IT, my daughter said, "Mom, that was an AI clone of my voice." She knew it immediately. Say, " I didn't know." So, there needs to be awareness about this, and there needs to be guardrails. We need to find out how to, how to protect those of us who don't know about this thank you thank you for letting me testify i appreciate it

oh we're sorry to uh over talk you guys i was gonna say

yes

you and your husband but

yes

and we're sorry this have ever happened to you so

thank you

next i'd like to introduce paul bender uh he's executive vice president for risk fraud and cyber security at the american bankers association where it leads the association's initiatives in fraud cyber security, physical security, and information security, practices, and chairs the aba fraud coordination group. Thank you for being here. Be, please begin your testimony.

Chairman Scott, ranking member Gill- Gillibrand, and members of the committee, thank you for the opportunity to testify, and I am so sorry what you went through. I think it proves the point why we're here today. We're all here to protect Americans from these, as you said, senator thugs, uh, these despicable people doing these things. The central point of my testimony is straightforward. Genitive AI is not replacing traditional scams, it is industrializing them. Genitive AI is making impersonation dramatically easier and more convincing. With very little technical skill, a criminal can mimic someone of victim trust and use personal information gathered online to make the scam feel authentic. What once required significant time, skill, and resources can now be done quickly, cheaply, and at scale. These crimes do not succeed because older Americans are unsophisticated. They succeed because professional criminals are very good at exploiting trust, fear urgency and authority ai simply gives them more powerful tools to do that a recent fbi warning shows how this threat is evolving the fbi reported that criminals are using ai generated videos of senior fbi officials fake social media profiles and spoofed government websites to impersonate the internet crime complaint center and target prior victims in other words the promise to recover stolen money becomes the mechanism for stealing even more bank impersonation scams are another serious example Just last week, the FCC warned they produced the highest losses of any impersonation scam category, and stressed one rule, banks will never ask you to move money to protect it. Additionally, a survey of fourteen large banks found identified bank impersonation scams rose a hundred and fifty percent from twenty twenty four to twenty twenty five. The threat is growing rapidly and much of the deception occurs before a bank ever sees the transaction. Banks are responding aggressively. They're using AI and advanced analytics to recognize behavior that may indicate and to intervene before money leaves an account. But technology is only part of the defense. A trained banker may notice that a longtime customer is suddenly trying to send a large payment while appearing frightened, secretive, or coached by someone on the phone. In those moments, human judgment and the relationship between banker and customer can be just as important as any algorithm. Education is equally critical. More than twenty-five hundred banks have participated in the ABA Foundation's Safe Banking for Seniors program. And the ABA and FBI have developed materials specifically addressing deepfake scams. We are also expanding trainers to help bank training to help bankers intervene when a customer has been drawn deeply into a scammer's story. For consumers, the answer is not to become experts at spotting deepfakes. The better approach is to change how we react to unexpected requests for money. Seniors should pause before acting, verify the story through a separate trusted channel. Families should consider establishing a private password for real emergencies. And remember, if it's a secret, it's a scam. Criminals depend on urgency in isolation. They do not want the victim talking to a family member, a banker, or law enforcement. We also need to recognize where the scam life cycle, uh, where in the scam life cycle banks enter the picture. We train consumers not to send money to someone they do not know and trust. The problem is that by the time they're ready to send the money, they often believe they know and trust the person on the other end. Generative AI is making it much easier for criminals to manufacture Increasingly accessible tools can alter a scammer's appearance in real-time video conversation to match the persona they are they are impersonating. AI can clone voices and translate conversations in real time, allowing criminals halfway around the world to communicate convincingly with an American victim. By the time the bank sees the payment, that relationship may have been developing for days, weeks, or even months through telecommunication networks, social media, online advertising, or messaging platforms. the victim is no longer sending money to a stranger. In their mind, they are sending money to someone they know and trust. Banks will continue to intervene where we can, but responsibility has to extend upstream to the telecommunications and on-line platforms where the impersonation begins and trust is manufactured. The best fraud prevention is to stop the criminal from reaching and deceiving the victim in the first place. My written testimony lays out a broad policy agenda. We need clear national leadership, which is why ABA is proposing a national office for scam and fraud prevention. We urge Congress to enact the scam act led by Senators Gallego and Marino, so no, so on-line platforms have meaningful obligations to verify advertisers, and rapidly respond to fraudulent ads. Stronger telecom safeguards are needed to keep criminals off calling networks and restore trust in caller ID. And federal support can help state and local authorities build specialized financial crime capacity. Older Americans should not be expected to distinguish on their own, and in real time, between a loved one and a cloned voice. or between their bank and a criminal using their bank's name and telephone number. Banks will continue investing, educating, intervening, and helping victims recover. But durable progress requires the part of the scam, every part of the scam ecosystem to help prevent the deception before the victims send money. Thank you, and I look forward to your questions.

Thank you, Mr. Bender. Next I'd like to introduce Matthew Ferraro. He's a partner in Crowell and Maureen's Privacy and Cybersecurity group, where he helps clients address complex regulatory matters. at the intersection of advanced technology, national security, and crisis management. Thank you for being here and please begin your testimony.

Chairman Scott, Ranking Member Gillibrand, and members of the committee. Thank you for the opportunity to appear before you today on this critical issue. My name is Matthew F. Ferraro. I'm a partner at the law firm Kroll & Morin, where I counsel on artificial intelligence, cyber security and regulation. I also previously served in government, most recently as senior counselor for cyber security. security and emerging technology to the secretary of homeland security, and previously as a US intelligence officer. I have been working on the policy and legal issues related to AI generated media or deep fakes since twenty nineteen. I should say I appear today in my personal capacity. My views do not represent those of my firm or any of my clients. Let me begin with a rhetorical question. How often today have you relied upon the voice of a loved one or the image of a trusted figure before sharing something of value. Perhaps your daughter called from college this morning to ask for a few hundred dollars and you sent it. Maybe you had a video call with a tech support worker and shared your account information. Tonight you may scroll through social media and see a video of a doctor endorsing an herbal supplement and decide to purchase it. Now consider, what if those representations were all fake? Not your daughter, not the technician, not a doctor. but voices and images forged by ai welcome to the world of deepfakes it is a world we have been living in for several years where ai can create realistic audio images and videos that appear true to life but are forgeries nearly anyone can access this technology often for free and their creations can be sent around the world instantly while ai generated media has legitimate creative uses deepfakes can supercharge scams and cyber frauds especially those targeting senior citizens. In the first quarter of twenty twenty five alone, financial losses from deepfake enabled fraud exceeded two hundred million dollars according to an industry report. Deloitte projects that generative AI could enable fraud losses to reach forty billion dollars in the United States alone by twenty twenty seven. Older adults bear a disproportionate burden. They are five times more likely to lose money in a scam than younger people. And according to the AARP, the top digital risk for seniors is scams and frauds. The forms these scams take are varied. For example, fraudsters use AI to generate fake ads featuring the likenesses of celebrities and trusted professionals to s- to sell a range of goods, from cryptocurrency to medical cures, as we've heard. Scammers have also cloned the voices of family members, telling their victims their loved ones have been kidnapped and demanding ransom. as Ms. Del Mastro testified so powerfully to a few minutes ago. And romance scammers can use AI-generated imagery to build emotional relationships with victims before uh stealing their money and their personal information. So what can we do about it? I offer three recommendations. First, we must prioritize education and awareness for all digital media consumers, especially seniors. We should deliver that education through trusted community channels, like senior centers, physicians, community groups, and houses of worship. We must cultivate the appropriate level of skepticism and of discernment. Given the deluge of AI-generated media that has and will come, the surest defender against deepfake dupes lies between our ears. Second, we should seek greater coordination across federal and state agencies and with industry. Existing laws already give regulators meaningful tools. The principle applies. If it is legal without AI, it is illegal with AI. Better knowledge sharing among policymakers, law enforcement, and industry can promote efficiency, improve threat mitigation, and bolster effective enforcement. And third, we should invest in technology. We should promote the use of both AI detection tools and provenance technology, that tags media as human-created, or ai generated. integrating those tools into our daily lives, much as we do with email spam filters, can help all americans, and especially seniors, tell facts from fakes. i wanna close on this note. we should not succumb to cynicism. new technologies, married to ancient vices, require responses. but they should neither slake our appetite for innovation and creativity, or extinguish our faith that we can, with the right steps, navigate the digital world with confidence. Thank you. I look forward to your questions.

Thank you. Now I'd like to recognize ranking members. Gillibrand to introduce the next witness.

Thank you, Chairman Scott. I want to move to introduce Ben Winters. Mister Winters is the Director of AI and Privacy at the Consumer Federation of America, where he leads CFA's advocacy efforts regarding data privacy and AI to advocate for consumers. Previously, Mister Winters worked as an Attorney Advisor for the Civil Rights Division of the Department of Justice and Senior Counsel at the Electronic Privacy Information Center. You may begin.

Thanks. Uh, good afternoon, Chair Scott, Ranking Member Gillibrand, and members of the committee. Uh, thanks for the opportunity to talk to you today. I'm Ben Winters, Director of AI and Privacy at the Consumer Federation of America, a non-profit membership organization serving consumers since nineteen sixty eight. online scam losses have been an unacceptable rate for years but have exploded since generative ai became commercially available and aggressively marketed in twenty thirty three losses reported to the fbi have ballooned from three and a half billion dollars in twenty nineteen to twelve and a half billion dollars in twenty twenty three to twenty one billion dollars last year in twenty twenty five again that is just to the fbi that is just who reported and that is just online scams cfa estimates in our true cost of scams report that the real number is closer to a hundred fifty billion dollars lost just last year with americans over sixty losing fifty five billion given this prevalence and scale it's not surprising that older americans ascribe scams is unavoidable and that it takes just one minute of distraction to fall victim generative ai is not the sole cause of rising scam losses but it is right now a scammer's dream it makes scams easier more effective and harder to trace This is just one part of multiple trend lines converging to uh this record losses. AI companies that aggressively push tools with no regard for harm, an executive branch that refuses to hold tech companies accountable, and confusion and uncertainty around healthcare eligibility, benefits, and economic security that scammers are exploiting in real time. Scammed content generated by AI systems are just one part of the puzzle. Unregulated and under-regulated technology is pervasive across the entire flow of how is a scam is created, targeted delivered and carried out cfa calls this the scam stack so this is everything from data brokers that sell detailed personal data enabling hyper-targeted scams based on sensitive characteristics like and this is a real example lists of people battling alzheimer's this is goes to mass communication methods like robotexters robocallers and social media companies like facebook uh through ads and feeds that allow obvious scam content to remain despite having the technology detect it This goes to payment platforms, banks, and crypto wallet providers that let the money go and allow obfuscation of the chain of custody. And lastly, reporting mechanisms on phones and social platforms that are ineffective, inconsistent, buggy, and disconnected from other reporting mechanisms like at the government agencies who can do enforcement. My written testimony goes into further detail about how generative AI enables elder thought at each level of that scam stack, uh, and the harm caused by its unchecked growth. But it's not just scams. AI systems like chatbots mislead by pretending to be real, sounding authoritative, trying to get you hooked, or giving dangerous medical or financial advice. And are also used by third parties to flood the information ecosystem with convincing falsehoods. That matters especially for older adults, isolated people, and those under health or financial stress. At scale, AI generated misinformation erodes trust in everything people see and hear. It's hard to avoid using these online platforms and scammy content scams and falsehoods fueled by ai are hurting us all i want to be clear that the devastating deception we're seeing is not a result of technological ineptitude for older users the exclusive domain of the dark web or overseas scam compounds but rather a crisis emboldened by the biggest tech companies we know paired with the failure to rein them in this is not an issue of personal responsibility but something congress is uniquely positioned to address my written testimony details more than a dozen legislative and oversight recommendations of specific bills that have already been introduced but i wanted to spend a few minutes talking about that today, uh a few of them, sorry. Uh so one, uh you should uh congress should hold platforms accountable for spreading, delivering and targeting scams by passing a bill like the scam act introduced by senators moreno and gallego. Should pass comprehensive data privacy laws, data minimization and bans on the sale of sensitive data, so it's no longer uh allowed to be able to sell uh lists of people battling alzheimer's for targeting. Uh if pass bills improving the reporting, and statistic status quo with a bill like report scams dot gov act by uh chair scott and uh senator hawson but would recommend improving that to require platform participation it is outrageous how complicated the answer is to what do i do when i've been scammed and reject any provisions that prohibit states from regulating technology or ones being pushed by tech companies right now to limit their liability legislation must be paired with sustained oversight of enforcement agencies like the ftc cfpb and fbi to ensure key authorities are focused on collaboration investigation and choking out the scam upstream not just chasing individual scammers after the fact there is no silver bullet for fixing this crisis but continued inaction only hurts americans thank you again for the opportunity to testify and i'm happy to answer any questions

thanks for thanks for being here so i bet your last job was interesting too

fascinating what they do ok senator tuberville would you like to start Thank you, Mr. Chairman. We got a mess. We know it. You know, technology bring us more and more every day. We've heard the problems. My my solution is education. At a young age, baby boomer boomers are in trouble because we're all gonna get scammed one way or another. And uh, not just domestically, but internationally. I mean, we're gonna It's big time. I hear a lot of problems with that. So, Mister Ferraro, how do we fix this? We got the problem. Give us a solution. I'm gonna ask everybody this, so get your solution ready.

Uh, Senator, there, there is no silver bullet as, as Mister Winters said, but I do think that education is very powerful. And if you're looking for an example, I would point the, to the committee, to the uh example of Finland, which has been very aggressive in teaching both young and old. this is in my written testimony, about uh both this information more broadly but synthetic media particularly, and they have found that it works, that you can actually instruct people to be on the lookout for false media, to as um as Mister Benda said, pause before you send any money, double check, uh do all the things that we should do before we make any rash decisions. And so I do think education is quite important. As I said, I think from a government perspective, inter-agency cooperation is key here um this is an area where there there isn't now a quarterback uh to to use a metaphor in the government on these issues and i think that is something that the committee might wanna consider i know that there are some bills pending for the committee on those issues and third as i mentioned in my testimony i think the technology piece is important as well i mean just draw the analysis or the analogy rather to email spam you you receive thousands of pieces of spam everyday you just don't happen to see them, because they're filtered out by spam filters. One could imagine a situation where similar technology operates in the background on your browser or on your phone to either tag or remove AI generated media uh and it that the technology largely exists now it's just a matter of integrating it. So I would I'll offer those.

Well, if you do anything in life, you know the big thing is when you get a manual, you get ethics with it first. And the same thing with, with AIAI, we have to teach our kids ethics. Again, a lot of us are lost and gone. But uh, our young kids can learn it. Mister uh, Benda, you uh, have worked for DHS and DOW. Uh, give us your solution.

I have, sir. So I, I think you're right, education is a key pillar. I think you start out young. Uh, I think having a national campaign on how to fight AI deepfakes, how to fight scams and fraud is important. ABA runs financial literacy programs for children. integrating the ai threat into that and scam threat into that is something that we're looking at doing uh i think the national campaign could address people of all ages and and i have hope for boomers sir i have some boomers that i love my mother-in-law and one particular um and i think i think the focus needs to be on also a safe banking for seniors so how do we get the message to them in in channels that they respect and they use uh whether it's directly from a banker or from people they trust and the other piece i would point to is we have to enable uh and stop the tech companies from allowing the impersonation that happens out there you're, you should be able to trust what's on your caller ID. If you put a name and number on there, the telecoms should be held accountable if it's not the right person calling. The social media platforms need to stop the impersonation scams that are out there. Uh, they shouldn't be allowed to post the videos that the good doctor talked about. So those types of things I think, uh, need more controls, and I think there's some legislative solutions that are out there.

Yeah. Miss Mastro, I'm, feel your pain, I, my mother-in-law got scammed, and this is internationally, she got a call from my granddaughter. She was in Europe.

Mm-hmm.

Please send ten thousand dollars. I need it bad. I'm broke. I gotta get out of this situation. She sent it. She sent it. She wasn't in trouble other than the fact she just needed the money.

Mm-hmm.

Sounded just like her. What kind of legislation you think we need to do for something like this? Have you talked to anybody since your debacle?

Um, I have not spoken to any legislation, uh, legislators, uh, until right now. Um, I have spoken to a, a few experts in in the field though, and I believe that, I mean, there are plenty of, there are plenty of courses and education out there at how to use AI and that is something you see on social media all the time, those ads, you know, how to learn how, like, take this class, learn how to use it. There's nothing out there on how to protect yourself from E AI. There's nothing out there. So that needs to happen. We need to know how to recognize it, and how to protect ourselves from it. Besides having a safe word, not answering the phone, you know, all the things that, that everybody immediately goes to. There's, there's got to be, there's got to be a, uh, an awareness that, that, uh, for senior citizens certainly. You know, all of us don't know how, you know, hand the remote to the kids to, to, you know, to do, to change things. I mean, we need to be aware. If I had been aware that my daughter's voice could possibly be an AI clone, I would have responded differently. But I did not know. So I think we need, uh, education is, uh, is key, is key, and awareness is key.

Just think about, we're just now beginning AI, just think of how much more they're gonna be able to fool people, and it's gonna get worse and worse. Thank you very much. Thank you, Mr. Chairman.

Thank you. there you go

thank you mr. chairman and thank you for having this uh hearing um as i think all of us know ai has a very real potential of changing everything about our lives uh it's changing how we work how we make decisions how entire industries are operating reshaping our economy energy systems national security It's affecting people's daily lives. And while it brings enormous opportunity, it also introduces risk that demands some serious oversight. And that's why we're here today. Uh, over the past year I've been thinking a lot about how we make I, AI work for all Americans, not just a few big companies. Uh, and that's what led me last year to put out a road map called AI for America. And the ideas behind the road map are pretty simple. And that is, if AI is gonna transform our economy, we need to have a real plan to make sure Americans aren't left behind. And that the technology is developed and used responsibly. And here's the thing, these systems are evolving very quickly. Even just quickly in the last few weeks. And there's a lot that we don't know, and there's a lot that the developers themselves don't know. about what some of these products are capable of. Older Americans need to be included in this conversation and not treated as just some afterthought as these products are developed and deployed. Um, about one in five Arizonans in my state are over the age of sixty-five. And when I hear from seniors in my state, they tell me they're concerned about what is real and what isn't real. Miss Del Mastro, you certainly have experienced that and i'm so sorry for what happened to you and your family um seniors are seeing more convincing phishing scams and ai generated scams and other forms of online fraud and sometimes people don't know if they're talking to a bot or a real person just yesterday i was on the phone with um my cable company i called and i was convinced that this was a artificial person. And I still think it may have been, it may have been augmented by a real person, and when I started questioning whether or not the person was real, it became the real person. So I think there are systems where one person, where you can have the AI overlaid and there's somebody monitoring, anyway. We sorted it out, I got the real person eventually. Um, so these are real fair concerns and it affects everybody, I share them. and i do spend in my former career as somebody who used a lot of technology um we so last week i introduced legislation with senator justice called the senior chatbot protection act this bill will require companies to clearly disclose when consumers are interacting with ai um and safeguard sensitive conversations and address deceptive design practices in their products that put people at risk We wanna help people make informed decisions while supporting not replacing human judgment and trusted relationships. So today, ranking member Gillibrand and I are asking the leading AI companies directly, how are you making sure your products are safe for older adults and are you designing products with them in mind? And do you even know how many older adults are using your products? And if we wanna make sure AI will help rather than harm, we need to know about, you know, who's using it, how they're using it, and how it reacts in the real world. So that's the purpose of my bipartisan AI, Aging with AI act, with, which the chairman mentioned, um, which we introduced, Chairman Scott and I in June. It will help us build the evidence we need by supporting research into how seniors use ai and its risks and benefits and we have a lot of work ahead of us ai companies have a responsibility to make their products safe uh to be honest about their limitations and protect their constituents who use them and congress has a responsibility to set clear rules demand accountability from tech companies and make decisions based on science data and facts Uh, we have an opportunity to get ahead of these problems. We often do not here in Congress, but we should, and we have to in this case, um, because the window is not gonna stay open forever on this. We gotta get this right now. So, in my remaining time, which I have none, uh, maybe the chairman could give me another forty-five seconds. Mister Winters, based on the harms you've seen and talked about during your testimony, how would the protections in the senior chatbot protection act make ai chatbots and voice assistants safer for older americans and is it important for congress to do this

yeah uh thank you senator for the question uh i think that some of the provisions in that bill are extraordinarily uh uh useful right as you mentioned uh it's nonsense to not be able to feel confident that you know that you're actually talking to a person or not right so that feels like an absolute baseline thing we can do uh and we can also sort of you know integrate significant uh restrictions on the way the chatbots are actually sort of looking to the user right uh to make it more obvious that it is a chatbot to uh uh make it clear what the choices from those companies are i think one thing not to take up too much time that i wanna underline is that the chatbot products especially are products they're they're tech companies have choices about what gets spit out by those uh chatbots how they are built and uh uh sort of how you interact them so i think that there are a number piece of legislation including the senior chatbot protection act uh that would be absolutely a positive step forward in in in in making sure those tech companies are pulling their weight in helping air

right thank you

thank you senator kelly um doctor amron you discovered a deepfake video using your face and voice to sell a product you never made and never endorsed can you walk through what you had to do to report it and what the process of trying to get it down uh was like

yeah it will it was not easy and it continued to go on um i'm gonna have to refer back to uh some notes i have by my team that got involved with this but um there was a lot of effort that went in to try to get this down um putting pressure on on meta uh to respond um and it took eleven days uh to get a response back from them uh the team uh never got transparency in terms of answers and any time line and things like that it was a very um non-transparent process with it um there were um there was no meaningful follow-up uh to get the facebook ads down uh the responses we got back were very generic um and uh you never actually knew that you were actually uh communicating with a real human being with this um so uh we eventually got at least the the meta ads uh down in uh eleven days the websites continued to be up there uh but then it reappeared uh weeks and months later again so it's just a perpetuating process