Altman flags mental health and cyber risks in AI hiring push

OpenAI chief executive Sam Altman has publicly acknowledged the expanding impact of artificial intelligence on mental health and cybersecurity, signalling a shift in how one of the world’s most influential AI developers frames the social risks of its technology. The remarks came alongside an announcement that OpenAI is seeking a Head of Preparedness, a role designed to anticipate and mitigate potential harms as AI systems grow more capable and widely deployed.

Altman said the new position would sit at the intersection of safety research, policy planning and operational readiness, with a remit that explicitly includes psychological wellbeing and digital security. By linking the hiring drive to these concerns, the OpenAI chief placed mental health and cyber resilience alongside more familiar debates around misinformation, automation and economic disruption.

The acknowledgement reflects mounting evidence that advanced AI tools are shaping how people think, work and interact online. Researchers have warned that prolonged exposure to conversational systems can influence emotional states, particularly when users rely on AI for companionship, advice or reassurance. Studies in psychology and human-computer interaction have highlighted risks ranging from emotional dependency to reinforcement of anxiety, especially when systems present confident responses without sufficient context or safeguards.

Altman’s comments also come amid intensifying scrutiny of AI-related cyber threats. As generative models become more accessible, security experts have reported their use in crafting phishing campaigns, automating social engineering and accelerating the discovery of software vulnerabilities. At the same time, AI is increasingly deployed defensively, helping organisations detect intrusions, analyse malware and manage complex threat landscapes. The dual-use nature of the technology has made preparedness a central concern for governments and companies alike.

The Head of Preparedness role is expected to coordinate internal safety teams with external stakeholders, including policymakers and academic researchers, to stress-test AI systems before deployment. According to people familiar with OpenAI’s internal discussions, the position will focus on scenario planning for high-impact risks, integrating insights from behavioural science, cybersecurity and ethics into product development. That approach reflects a broader industry trend toward formalised “red-teaming” and risk assessment as models approach higher levels of autonomy.

OpenAI has previously spoken about existential and systemic risks from AI, but the explicit reference to mental health marks a notable broadening of focus. Mental health professionals have increasingly called for clearer guidelines on how AI systems should handle sensitive conversations, particularly around self-harm, trauma and emotional distress. Some regulators are examining whether existing consumer protection and health frameworks are sufficient to address AI-mediated interactions that blur the line between tool and confidant.

In the cybersecurity domain, Altman’s remarks align with warnings from international agencies that generative AI could lower the barrier to entry for cybercrime. Law enforcement bodies have reported a rise in AI-assisted scams that combine convincing language with personal data harvested from breaches. At the same time, large technology firms argue that responsible deployment, access controls and continuous monitoring can limit misuse while preserving innovation.

Industry analysts say OpenAI’s hiring move underscores a recognition that technical advances alone are insufficient to manage AI’s societal footprint. “Preparedness is becoming as much about human impact as it is about technical safety,” said one AI governance specialist, noting that public trust increasingly hinges on how companies address psychological and security concerns. The creation of a senior role dedicated to these issues suggests an attempt to institutionalise that responsibility.

The announcement arrives as competition in the AI sector intensifies, with rival firms racing to release more powerful models for consumers and enterprises. This rapid pace has fuelled calls from academics and civil society groups for stronger oversight and clearer accountability. Some governments are drafting rules that would require companies to demonstrate risk mitigation strategies before deploying high-impact AI systems, particularly those used in healthcare, education and public services.