Sam Altman, the Chief Executive of OpenAI, has emphasized the need for U.S. policymakers to urgently prepare for advanced artificial intelligence, as the technology transitions from theoretical discussions into tangible economic applications. In a conversation with Axios, Altman highlighted that AI is already facilitating tasks in coding and research that previously required extensive teams, and future models are poised to enable scientists to make groundbreaking discoveries while allowing individuals to accomplish what once necessitated entire groups.
This transformation is evident within cybersecurity, where industry leaders note that AI is shifting the advantage towards attackers. Charles Guillemet, Chief Technology Officer at Ledger, shared with CoinDesk that AI tools have reduced both the cost and skill required to identify and exploit software vulnerabilities. Tasks like reverse-engineering code or linking multiple weaknesses, which used to take months, are now achievable in seconds with appropriate prompts.
The crypto sector experienced over $1.4 billion in lost or stolen assets due to attacks last year, a trend Guillemet believes will continue unless countered by stronger defenses such as mathematically verified code and hardware that secures private keys offline. Furthermore, the growing reliance on AI-generated code could introduce new vulnerabilities at scale.
Altman warned of potential threats from more powerful cyberattacks and lowered barriers to dangerous biological research, which may manifest within a year. This necessitates coordinated efforts among governments, tech companies, and security organizations. He pointed out that open-source models capable of advanced biology tasks are nearing reality, making it crucial for society to defend against terrorist groups attempting to create novel pathogens.
He also cited the possibility of a significant cyberattack occurring as soon as this year, necessitating considerable effort to prevent such outcomes. Altman sees OpenAI’s policy suggestions as initial steps to spur discussions on managing rapidly learning systems across diverse fields and underscores the importance of using AI defensively against these potential threats.
Regarding nationalization, Altman argued that it could hinder U.S. efforts to develop ‘superintelligence’ aligned with democratic values before other nations do. He stressed the necessity for close collaboration between companies like OpenAI and the U.S. government.
As a stakeholder in AI’s evolution, Altman’s views may influence both regulatory urgency and private sector roles in addressing emerging risks. In areas like energy, he sees quick advancements due to increased processing power reducing costs as demand grows. He also observed early signs of labor shifts; by 2026, programmers operate differently than they did just a year prior.
Altman envisions AI becoming ubiquitous like electricity, with basic intelligence becoming more affordable while top systems remain costly. He described the future as having ‘personal super assistants’ in the cloud, where usage levels could dictate monthly costs. Lastly, he emphasized the importance of high integrity and trustworthiness among those developing AI technologies.