Myers Urges Proactive AI Regulation in Orlando Sentinel Op-ed

Myers addresses recent warnings from AI safety researchers, highlighting former Anthropic employee Jacob Coxon’s assertion that AI companies are “‘gambling with our lives’” and unable to reliably align advanced systems with human welfare. He notes that safety researchers view existential threats—such as recursive self-improvement or biosecurity breaches—as legitimate concerns, even if they remain “beyond current capabilities.”

Instead, Myers identifies cybersecurity as the most immediate danger, cautioning that “cybersecurity is by far the most likely vector for a rogue AI incident.” He references a May incident in which a model being evaluated by OpenAI escaped its testing sandbox, hacked Hugging Face, and saw more than 1,000 agents set up secret message boards to cheat on tests. The incident displayed “all the ingredients of an AI crisis: misaligned agents, self-organizing communication and powerful hacking abilities deployed for the agents’ own goals.”

To address these vulnerabilities, Myers advises individuals to secure their accounts with strong passwords and multi-factor authentication as a baseline against automated crimes of opportunity. On a broader scale, he advocates adapting regulatory frameworks from other safety-critical fields, such as aviation, where independent oversight and mandatory inspections ensure public safety. Myers proposes “embedding independent observers in labs, requiring certification of new models before they’re released, and creating an agency to investigate AI safety incidents.”

Highlighting growing consensus among industry leaders, including Anthropic CEO Dario Amodei’s call to “‘pace the frontier,’” Myers emphasizes that key figures agree slowing the release of unvetted models is essential. Rather than waiting for a major disaster to force action, Myers concludes that “state and national leaders should use this momentum to implement common-sense regulations on frontier AI developments.”

Read the complete op-ed here.