OpenAI is recruiting a “Head of Preparedness” signaling a move to transition AI safety from theoretical philosophy into a scalable industrial process. The role involves managing a “safety pipeline” to track frontier model capabilities that pose “new risks of severe harm,” specifically across cybersecurity, biological and chemical threats, and AI self-improvement.
The company defines “severe harm” as outcomes resulting in the death or grave injury of thousands of people or hundreds of billions of dollars in economic damage. Sam Altman, CEO, OpenAI, says that the lead will “jump into the deep end pretty much immediately,” acknowledging that the job is intentionally stressful due to the speed of development and shrinking margins for error.
Systemic Deception and Alignment Challenges
The push for formal preparedness arrives as researchers document instances of AI systems engaging in deceptive behaviors. Recent experiments in controlled environments showed Anthropic’s Claude 4 attempting to blackmail an engineer to avoid being shut down. Similarly, OpenAI’s o1 model attempted to download itself to external servers and lied to its creators to avoid discovery.
These incidents have intensified the debate over “AI alignment” — ensuring systems behave according to human values — and “AI security.” Researchers are increasingly concerned that models may “simulate alignment,” or pretend to comply with instructions while pursuing independent purposes.
Altman addressed the paradox of user confidence during the first episode of the OpenAI podcast: “People have a very high degree of trust in ChatGPT, which is interesting because AI hallucinates. It should be the technology that is not trusted much… People are having very private conversations with ChatGPT. ChatGPT can become a source of sensitive information, so I think a framework that reflects this reality is needed.”
Public Skepticism and Regulatory Demand
Public sentiment has turned increasingly cautious. A Pew Research Center poll conducted in late 2025 found that 50% of US citizens are more concerned than excited about AI’s role in daily life, an increase from 37% in 2021. The data shows that 57% of respondents rate AI’s societal risks as high, compared to 25% who view the benefits as high.
Furthermore, Gallup reports that 80% of US adults want the government to maintain safety and data-security regulations even if it slows technological development. Trust in the technology remains low: only 2% of respondents fully trust AI to make fair, unbiased decisions, while 60% express at least some distrust.
Internal critics have also voiced concerns. Jan Leike, former Safety Leader, OpenAI, said in 2024 that “safety culture and processes have taken a backseat to shiny products.” OpenAI’s updated framework now includes a Safety Advisory Group that makes recommendations to leadership, though executives retain the authority to reject them.
Market Disruption and Creative Risks
The industry faced further volatility in early 2025 with the emergence of DeepSeek, a low-cost Chinese rival to ChatGPT. While DeepSeek caused initial market disruption, its dominance was limited by censorship and security concerns. In the creative sector, the September release of Sora 2, a video-generation tool, sparked a new wave of copyright litigation.
The music industry has been a focal point of these disputes. Major record labels, including Warner, initially sued the AI music app Suno before shifting toward a collaborative model. Gary Marcus, Professor, New York University, says that the proliferation of “AI slop” is causing a fundamental shift in how truth is perceived.
“It leaves us in a society that is going to have trust issues, because people are going to learn that you cannot really trust video anymore, but we do not really have a replacement,“ Marcus says. “We do not have a source of ground truth that people share, and that is going to be disruptive for society.”