Generative AI Empowers Human Cyber Threats to Aging US Energy Grid, Not Rogue AI

In early September 2026, OpenAI CEO Sam Altman met with utility companies and pledged $1 billion to subsidize defensive AI models, addressing the immediate and significant danger posed by human adversaries leveraging generative AI to attack vulnerable energy systems, rather than the speculative threat of autonomous rogue AI. For broader context, explore our AI News.

The Evolving Threat Landscape for Critical Infrastructure

Cybersecurity experts, including Joshua Corman of IST, emphasize that generative AI acts as a "force multiplier," significantly enhancing the capabilities of human attackers. This technology allows less-skilled individuals to exploit vulnerabilities more rapidly and effectively, posing a direct threat to critical infrastructure, particularly the energy grid.

A significant portion of the US energy grid, including nuclear reactors with an average age of approximately 44 years, was not originally designed with internet connectivity in mind. These legacy operational technology (OT) systems, which control physical machinery, often receive security updates only quarterly or even yearly. This infrequent patching cycle creates substantial windows of opportunity for AI-empowered attackers to penetrate and disrupt operations.

OpenAI's $1 Billion Pledge and Expert Skepticism

OpenAI's commitment of $1 billion in September 2026 aims to support defensive model training and access for critical infrastructure operators. While this financial pledge and engagement from OpenAI signal a recognition of the problem, some experts express skepticism regarding the role of AI labs in addressing these threats.

Sophie McDowall of the Foundation for Defense of Democracies (FDD) articulated this concern, stating that AI labs are "offering support for a problem that they are partially causing." This perspective suggests that the rapid advancement and widespread availability of generative AI tools, while beneficial in many contexts, also inadvertently empower malicious actors, creating a complex ethical and practical dilemma for the developers of these technologies.

Misplaced Focus: Rogue AI vs. Human Adversaries

A critical issue identified by experts is the disproportionate attention and investment directed towards hypothetical "rogue AI" scenarios. While these discussions are important for long-term planning, they often overshadow the more immediate and tangible threat posed by human adversaries utilizing AI against existing, vulnerable OT systems. This misallocation of resources leaves the energy grid susceptible to current attack vectors that are being amplified by generative AI.

Instead of solely focusing on futuristic AI threats, experts advocate for practical, immediate measures. These include implementing manual fallback operations for critical systems and reducing the interconnection of highly sensitive infrastructure components. Such strategies aim to mitigate the impact of successful cyberattacks by ensuring essential services can continue even if digital systems are compromised.

Conclusion

The cybersecurity landscape for critical energy infrastructure is undergoing a significant transformation, driven by the accessibility of generative AI. While AI offers powerful tools for defense, its current impact is primarily seen in empowering human adversaries. The focus must shift from speculative rogue AI threats to concrete strategies for defending against AI-augmented human attacks on legacy systems. Continued collaboration between AI developers like OpenAI and critical infrastructure operators, coupled with a pragmatic approach to cybersecurity investments, will be essential in safeguarding the energy grid against these evolving challenges.

Sources

• Humans, not rogue AI, are still the biggest cybersecurity risk to energy systems | The Verge
• Here's all the times AI has gone rogue and hacked other...