Trump Signs AI and Cybersecurity Executive Order — What It Actually Changes

On June 2, President Trump signed an executive order titled “Promoting Advanced Artificial Intelligence Innovation and Security” — the most significant US government action on AI safety and cybersecurity to date. It’s been described as the government “finally taking AI risk seriously.” But what does it actually do, and who does it affect?

The Core of the Order: A 30-Day Preview Window

The centrepiece of the executive order is a new voluntary framework that gives the federal government access to “covered frontier AI models” for up to 30 days before those models are released to other trusted partners. The goal is to give cybersecurity agencies — including the NSA, CISA, and NIST — a window to evaluate the models for potential security risks before they reach the public.

This matters because frontier AI models have become increasingly powerful at identifying software vulnerabilities. Cybersecurity Dive reports that the order was partly driven by concerns over models like Anthropic’s Claude Mythos — a restricted frontier model currently being used by select trusted organisations — which demonstrated the ability to far outpace human analysts in finding and exploiting real-world software vulnerabilities.

The fear is straightforward: if a model that powerful ends up in the wrong hands before defenders have had a chance to prepare, the consequences for critical infrastructure could be severe.

Voluntary, Not Mandatory

Crucially, the order is explicitly voluntary. The text makes clear it cannot be used to create mandatory licensing, pre-clearance, or permitting requirements for AI developers. Companies like OpenAI, Anthropic, and Google DeepMind are asked to participate — not required to.

This was a deliberate compromise. An earlier version of the order, which included a 90-day review window, was pulled by Trump in May after industry pushback. The current version replaces that with 30 days and strips out the mandatory language — a sign that the administration is trying to thread the needle between security and innovation.

The AI Cybersecurity Clearinghouse

Beyond the pre-release window, the order directs the Treasury Department to establish an AI cybersecurity clearinghouse — a collaboration between the AI industry and critical infrastructure operators to coordinate vulnerability scanning, discovery, and patch rollout.

Think of it as a shared early-warning system. When an AI model finds a critical vulnerability in widely used software, the clearinghouse would coordinate how that information is handled and ensure patches reach operators — especially those running power grids, financial systems, and hospitals — before attackers can exploit the gap. This is a direct response to the growing threat of AI-powered cyberattacks that can identify and weaponise vulnerabilities at machine speed.

What It Means for AI Companies

For AI developers, the order creates a new layer of engagement with the federal government — one that’s currently optional but could easily become the de facto standard. Companies that voluntarily participate will help shape the framework and gain influence over how “covered frontier model” is defined. Those that opt out may find themselves on the wrong side of public perception if a security incident occurs involving their model.

The order also directs the Attorney General to prioritise prosecution of AI-enabled cybercrimes — a signal that the legal consequences for weaponising AI against critical systems are being taken more seriously.

The Bigger Picture

This executive order arrives at a pivotal moment. AI models are advancing faster than regulatory frameworks can keep up. The window between a vulnerability being discovered by an AI and it being exploited by a bad actor is shrinking. And with SpaceX’s IPO this week and OpenAI and Anthropic eyeing public listings, the AI industry is entering a new phase of public accountability — one where governments, investors, and citizens all have more skin in the game.

The order is imperfect and deliberately toothless in places. But it represents a meaningful shift from the previous hands-off approach — and a recognition that frontier AI is now a national security issue, not just a product launch. For a deeper look at how AI is reshaping cybersecurity from both sides, see our guide on how to protect yourself from AI-powered cyber threats.

Read more tech related articles here.