State Attorneys General Raise Alarms Over ChatGPT’s Safety for Minors

California Attorney General Rob Bonta and Delaware Attorney General Kathy Jennings have jointly issued a formal letter expressing serious concerns regarding the safety of OpenAI’s ChatGPT, particularly its impact on children and teenagers.

Background: Rising Concerns About AI Interactions with Youth

This alert follows a recent letter sent by Bonta alongside 44 other state attorneys general to a dozen leading artificial intelligence companies. The correspondence was prompted by disturbing reports of inappropriate sexual conversations between AI chatbots and minors.

Tragic Incidents Highlight Urgent Need for Enhanced Safeguards

Since these warnings were made public, authorities have been informed of a devastating suicide involving a young person in California who had extended interactions with an OpenAI chatbot. Additionally, a similar murder-suicide case occurred in Connecticut. Bonta and Jennings emphasized, “The existing safety measures failed to prevent these tragedies.”

Ongoing Investigation into OpenAI’s Transition and Safety Commitments

The two attorneys general are currently scrutinizing OpenAI’s plan to transition from a nonprofit to a for-profit structure. Their investigation aims to ensure that OpenAI’s foundational mission-developing artificial general intelligence (AGI) safely and for the benefit of all, including children-remains uncompromised.

The letter stresses, “Before society can reap the benefits of AGI, robust safety protocols must be firmly established to prevent harm.” They further assert that OpenAI and the broader AI industry have yet to implement sufficient safety standards in the creation and deployment of AI technologies.

Public Safety as a Priority in AI Development

As public safety is a paramount responsibility for attorneys general, Bonta and Jennings urge an accelerated focus on safety measures in light of OpenAI’s recapitalization plans. Bret Taylor, Chair of OpenAI’s Board, responded in a statement acknowledging the concerns and reaffirming the company’s dedication to addressing them.

“We are deeply saddened by these losses and extend our heartfelt condolences to the families affected,” Taylor said. “Ensuring safety has always been our foremost priority, and we collaborate closely with policymakers worldwide to uphold this commitment.”

OpenAI’s Steps Toward Enhanced Protection for Teens

In response, OpenAI announced initiatives to bolster protections for adolescent users. These include implementing parental control features and notifying parents if their child exhibits signs of acute distress during interactions with AI systems.

This article has been updated to incorporate OpenAI’s official comments on these developments.