"The formalization of government pre-release reviews is marking the end of AI’s ‘Wild West’ era," Ram Varadarajan, CEO at Acalvio, said, summing up a central theme running through security commentary after the White House moved to inspect advanced models before general release.
Executive order: pre-release model reviews and the 90-day proposal
President Trump has signed an executive order requesting AI companies to show models to the federal government so the government can assess their capabilities prior to a full release. Several security leaders in the field endorsed the idea of pre-release scrutiny but emphasized that a single checkpoint is not the whole solution. Diana Kelley, Chief Information Security Officer at Noma Security, described a "90-day government review" as potentially useful as one checkpoint while warning that model safety evaluation is "complex and ongoing"—especially after models are deployed and connected to agents, code execution, enterprise data, identity systems or critical infrastructure workflows.
Ram Varadarajan: transforming frontier AI into a regulated strategic industry
Varadarajan argued the executive order marks a transition point: "Geopolitical alignment and national security clearances are going to become as critical to a frontier lab’s valuation as its raw compute." He framed the move as the end of an era in which developers could iterate without formal government review, saying this change will shift "frontier AI from a pure-play tech bet into a regulated strategic industry." Robert Costello, Chief Digital and Information Officer at Merlin Group, called the review period "a tremendously positive step," noting it gives the federal government a window to assess releases and work with industry before concerns become problems.
Operational deployment risks: Marcus Fowler, Diana Kelley, and John Gallagher
Several leaders cautioned that pre-release review must be complemented by operational security work. Marcus Fowler, CEO of Darktrace Federal, said the challenge is "ensuring AI systems are deployed securely once they move into real operational environments" and pointed to the need for "clearer visibility into how those systems behave" and guidance from NIST’s AI Agent Standards Initiative and forthcoming CISA guidance. Diana Kelley added that review should account for what a model can access, its autonomy, and whether guardrails are enforced in production. John Gallagher, Vice President at Viakoo, warned that advanced AI "is a massive supply chain risk" for operational technology and physical security systems, and praised efforts such as Project Glasswing and testing "against OT systems" to establish use-case specific guardrails.
Speed gap and proactive security: Dave Gerry, Randolph Barr, and industry measures
Industry executives repeatedly returned to the problem of speed. Dave Gerry, CEO at Bugcrowd, said attackers now "utiliz[e] AI to accelerate their pace and frequency of attacks," and that "adversaries move faster than the government"—a reality he said forces public agencies to play catch-up. Gerry urged a shift from compliance-and-post-incident approaches to proactive vulnerability discovery, noting that bug bounty and vulnerability disclosure programs have been effective in some federal contexts but are not yet standard across agencies or critical infrastructure operators. Randolph Barr, Chief Information Security Officer at Cequence Security, quantified the change in kind of risks: "About two-thirds of current AI-related incidents still originate from traditional weaknesses, however, the remaining third are uniquely 'AI-native'—model and data poisoning, prompt injection, and autonomous agents that can chain together API calls."
Legal framework and market access: Collin Hogue-Spears and Rajeev Gupta
Collin Hogue-Spears, Senior Director of Solution Management at Black Duck, framed the executive order as a voluntary lane for pre-release scrutiny: "Voluntary is not the policy floor. It is the legal ceiling on executive AI review without Congress." He noted the administration already sent Congress a March 2026 AI legislative framework and that the Center for AI Standards and Innovation (CAISI), a division of NIST, had voluntary testing agreements with Google, Microsoft, and xAI before the order. Hogue-Spears contrasted the U.S. approach with other jurisdictions, pointing out that China required generative-AI service filings in 2023 and that the European Union made general-purpose AI documentation and cooperation obligations applicable in August 2025 under the AI Act. Rajeev Gupta, Co-Founder & CPO at Cowbell, said the government "simply isn’t equipped to meaningfully oversee frontier AI models on its own" and recommended a public-private consortium—citing historical precedent in the nuclear sector—as a better model for shared technical capacity and accountability.
What this means for AI labs, federal agencies, and state and local governments
- AI labs: Expect pre-release scrutiny to factor into valuations and market access; some labs already have voluntary testing agreements with CAISI, and review could influence commercial rollout strategies.
- Federal agencies: Will need to clarify technical authority and staffing to evaluate models at scale; agencies such as NIST and CISA are cited as important sources of standards and operational guidance.
- State and local governments: May continue to lag in capacity and funding, a concern raised by Dave Gerry, who noted cybercriminals target smaller, less sophisticated organizations disproportionately.
Security leaders broadly endorse the idea of pre-release review but are unanimous that a voluntary, single-window check is only the start: operational monitoring, independent testing, clear accountability channels, and faster, proactive security practices will be required to keep pace with both traditional and "AI-native" threats. The executive order sets a new expectation that advanced models will be visible to federal reviewers before release; whether that expectation becomes a binding, enforceable framework or remains a voluntary lane will depend on the legislative and institutional choices that follow.
