Kelly, Gillibrand Press AI Companies to Strengthen Protections for Seniors

This week, Senators Mark Kelly (D-AZ) and Kirsten Gillibrand (D-NY), members of the Senate Special Committee on Aging, pressed the CEOs of nine AI companies for information about the safeguards they deploy on their AI chatbot companion products for at-risk users like older adults, people with disabilities, and individuals in crisis. “AI products may help older adults detect health problems and live independently. Generative AI may benefit people with disabilities by helping them communicate with others or operate assistive technology. Caregivers may find generative AI helpful for coordinating medical appointments or tracking the health of a loved-one,” wrote the senators. “However, to live up to this potential, generative AI companions must be safe to use, incorporating the needs of older adults, people with disabilities, and their caregivers. “People with disabilities may be particularly vulnerable to fraud facilitated by generative AI for a variety of reasons, including but not limited to reliance on caregivers, limited transportation options, limited access to sign language interpreters and assistive devices, and isolation from the community,” continued the senators. “Design processes that do not allow feedback from older adults and people with disabilities suggest that these risks may not be addressed on the front end of product development, leaving these populations vulnerable when using AI tools.” According to the FBI, seniors lost nearly $8 billion to scammers in 2025, with an average loss of $39,000 per person. This includes more than $350 million lost in AI-related scams. Total losses increased by 59 percent from 2024, due in part to the increasing sophistication of AI technology. Common AI scams include phishing attacks, where fraudsters trick individuals into revealing sensitive information; voice cloning, where generative AI is used to mimic a familiar voice; and image creation scams, where generative AI is used to forge documents and identification to make fraudulent messages appear trustworthy. Specifically, Kelly and Gillibrand asked the CEOs of OpenAI, Microsoft, Google, Anthropic, Perplexity AI, X, Meta, Luka Inc., and Character Technologies Inc. to provide information on: What safety testing these companies perform for their AI companion products How the company tracks adverse events/harms involving their AI products, including hospitalizations, financial losses, or deaths What types of safeguards are built into their AI companion products to mitigate harm and protect at-risk users like seniors How the company includes older adults, people with disabilities, caregivers, and advocacy groups in the development of their AI products Whether their companies provide independent researchers with access to data or tools to study the effects of their AI products on seniors and people with disabilities In June, Kelly introduced the bipartisan Aging with AI Act , which would require research about how AI tools impact older Americans. The legislation was followed up by the introduction of Kelly’s bipartisan Senior Chatbot Protection Act , a bill that would keep older Americans safe by requiring AI chatbot and voice assistant companies to provide clear disclosures, safeguard sensitive conversations, and end deceptive design practices in their products. Click here to read the full text of the letter.