AI Company Bankruptcies Pose Existential Threat to User Data
תומצת מתוך Calcalist על ידי baba
The potential bankruptcy of AI companies like Character.AI and OpenAI poses a significant risk to sensitive user data, including private conversations and trade secrets. Attorneys general are seeking to protect user information, especially from minors, fearing data could be sold as assets during insolvency proceedings. Experts warn that these data repositories are also strategic assets in the global AI race, raising national security concerns. It is proposed that AI companies, like financial institutions, should develop "living wills" to manage user data in the event of bankruptcy.
הידיעה ב-6 שורות · תקציר של baba
- AI company bankruptcies raise concerns over the fate of sensitive user data, including private conversations.
- Character.AI filed for bankruptcy protection, prompting calls to safeguard user data, especially from minors.
- Attorneys general from 40 U.S. states are demanding protection for user data in AI company insolvencies.
- The 23andMe bankruptcy case highlights the risk of sensitive data being sold during insolvency proceedings.
- AI data repositories are considered strategic assets in the global competition between the U.S. and China.
- Experts suggest AI companies need "living wills" to manage user data during bankruptcy.
The potential bankruptcy of artificial intelligence companies raises serious concerns about the fate of sensitive user data, including private conversations, medical information, and trade secrets. This issue has moved from theoretical to practical with Character.AI, a chatbot company, filing for bankruptcy protection in the U.S. In response, attorneys general from 40 U.S. states have petitioned the court to safeguard user data, particularly that of minors, fearing that intimate conversations could be sold as assets. Simultaneously, eight U.S. states have filed lawsuits against OpenAI regarding child safety. The rising cost of insurance against tech company defaults in the AI sector further highlights the vulnerability of these firms.
The case of genetic testing company 23andMe, which filed for bankruptcy in 2025 and had its sensitive genetic databases sold for $305 million, serves as a stark warning. Regulatory intervention was required to enforce privacy protections on the buyer. Experts question what will prevent a similar scenario where millions of intimate AI conversations become saleable assets.
Professor Nitzan Galslevich Pakin, who previously researched "too big to fail 2.0" in digital service providers, argues that AI companies represent a third, more troubling generation of this risk. Unlike social media, where users intentionally share information, chatbot interactions are perceived as private. This data, if a company collapses, may not disappear but could be sold, transferred, or used for training other systems, posing risks beyond privacy.
In the context of U.S.-China AI competition, these data repositories are also strategic assets. They can contain trade secrets, business plans, code, and insights into millions of users' behaviors. Transferring such data to foreign entities or those with ties to rival nations could create national security risks, enabling espionage, technological advantage, or public manipulation.
Given Israel's rapid adoption of international AI systems and the sensitive data entrusted to them, these risks warrant serious consideration. Drawing a parallel to financial institutions, which are required to have "living wills" for orderly collapse, it is suggested that major AI companies should also develop plans for user data in cases of insolvency. Such plans should include restrictions on sensitive data sales, protections for minors' conversations, data deletion options, and oversight of transfers to foreign buyers. Boards of directors must also integrate the potential collapse and data fate into their risk management strategies, emphasizing proactive planning over reactive crisis management.