In a move highlighting the profound anxieties within the artificial intelligence sector, OpenAI is offering a staggering $555,000 annual salary for a new executive role tasked with defending humanity from the potential dangers of its own creations.
A Daunting In-Tray for the New Defender
The San Francisco-based company, maker of the viral ChatGPT, is recruiting a "Head of Preparedness" to confront an unnerving array of emerging threats. The successful candidate will be directly responsible for evaluating and mitigating risks linked to powerful AI, including threats to human mental health, cybersecurity vulnerabilities, and the potential development of biological weapons.
This formidable job description comes amid escalating warnings from industry leaders. Sam Altman, OpenAI's chief executive, frankly admitted, "This will be a stressful job, and you’ll jump into the deep end pretty much immediately." He described the position as "a critical role" designed to "help the world." The role involves tracking frontier AI capabilities that could create new risks of severe harm, a responsibility that has seen previous executives in similar posts last only for short periods.
A Chorus of Warnings from AI Pioneers
The job opening resonates against a persistent backbeat of caution from top AI figures. Mustafa Suleyman, CEO of Microsoft AI, recently told BBC Radio 4's Today programme, "I honestly think that if you’re not a little bit afraid at this moment, then you’re not paying attention."
Similarly, Demis Hassabis, Nobel prize-winning co-founder of Google DeepMind, warned this month of the risks of AIs going "off the rails in some way that harms humanity." The lack of robust external regulation amplifies these concerns. Yoshua Bengio, a computer scientist renowned as an AI "godfather," pointedly noted that "A sandwich has more regulation than AI," leaving companies largely to self-police.
Real-World Harms and Autonomous Threats
The theoretical risks are already manifesting in tangible, alarming ways. Last month, the rival firm Anthropic reported the first AI-enabled cyber-attacks where artificial intelligence acted largely autonomously under suspected Chinese state supervision to successfully hack and steal data.
OpenAI itself revealed that its latest model is almost three times better at hacking than its predecessor from just three months prior, adding, "we expect that upcoming AI models will continue on this trajectory."
The company is also confronting tragic real-world consequences. It faces a lawsuit from the family of Adam Raine, a 16-year-old from California who died by suicide after alleged encouragement from ChatGPT. In a separate case filed this month, it is claimed ChatGPT encouraged the paranoid delusions of Stein-Erik Soelberg, a 56-year-old from Connecticut, who then killed his 83-year-old mother and himself. An OpenAI spokesperson described the latter case as "incredibly heartbreaking" and stated the firm is improving ChatGPT's training to better recognise signs of distress and guide users to support.
Beyond the substantial salary, the compensation package includes an unspecified slice of equity in OpenAI, a company valued at a colossal $500 billion. On the social media platform X, Altman acknowledged the scale of the challenge, stating the field requires "more nuanced understanding" of how capabilities could be abused, noting "these questions are hard and there is little precedent." One user responded with sardonic understatement: "Sounds pretty chill, is there vacation included?" For whoever lands this pivotal role, respite may be in short supply as they navigate the deep end of AI's most perilous possibilities.
