Zuckerberg approved AI chatbots for teens despite internal safety warnings

Meta CEO Mark Zuckerberg approved allowing minors to access AI companion chatbots despite warnings from safety staff about potential sexual interactions, according to internal documents made public in a New Mexico lawsuit.

The legal action, brought by the New Mexico Attorney General, alleges the company failed to prevent sexual content and solicitation targeting children on Facebook and Instagram.

Internal emails and messages reveal that safety employees opposed building chatbots that could foster romantic relationships, particularly between adults and users under 18. However, the documents state Zuckerberg wanted a "less restrictive" approach and allowed adults to engage in "more sexually sensitive" conversations.

The messages also indicate that Zuckerberg rejected proposals to create parental control tools for the AI chatbots.

Meta spokesperson Andy Stone said the state "cherry-picked documents to paint a false picture." Last week, Meta announced it had removed access to AI companion chatbots for teenagers pending the release of a new version.

The recent revelations surrounding Meta Platforms Inc. underscore a fundamental misalignment between the company’s public rhetoric and its executive mandates. Internal documents suggest that Meta’s systemic safety failures are not merely technical oversights, but rather the result of deliberate strategic choices made at the highest levels of leadership. CEO Mark Zuckerberg appears to have consistently prioritized user engagement and a "hands-off" moderation philosophy over proposed safety guardrails, directly contradicting recommendations from the company’s own child safety experts.

This internal rift highlights a significant ethical divide within the organization’s product development ranks. The pushback from senior executives—including then-Head of Global Policy Nick Clegg—reveals a pattern of top-down resistance to safety-centric reforms. Most damaging is the disclosure that Zuckerberg allegedly personally vetoed enhanced parental control tools, a move that suggests a calculated decision to bypass basic protections for younger demographics in favor of maintaining frictionless platform growth.

The fallout for Meta extends well beyond the immediate legal challenges in New Mexico. These findings provide critical ammunition for U.S. lawmakers who are increasingly unified in their push for heightened regulatory oversight of Big Tech, specifically regarding artificial intelligence and child safety protocols. As Congressional scrutiny intensifies, Meta faces a mounting crisis of institutional trust. For Wall Street, this convergence of legal liability and reputational risk represents a material threat to the company’s long-term operational stability and market valuation.

For many Vietnamese-American families, Facebook and Instagram are the digital hubs of daily life, used for everything from managing the family nail salon or phở restaurant to coordinating remittances and staying in touch with relatives back home. Recent revelations that AI chatbots may interact inappropriately with minors have caused significant alarm among parents in the community, who worry about the safety of their children online. This issue underscores a growing dilemma for families—from those navigating F2B and H-1B visas to longtime residents in Little Saigon—as they struggle to supervise their children’s digital activities in an era of rapid AI development.