President Trump has scrapped his own executive order requiring artificial intelligence companies to voluntarily preview advanced AI models with national security officials before public release, according to sources familiar with the decision.
The reversal marks a significant victory for Silicon Valley's most influential tech leaders, who mounted aggressive opposition to even the modest oversight measure. The decision underscores a troubling dynamic in technology policy: industry lobbying power has successfully blocked a light-touch regulatory framework designed to address national security concerns.
The Abandoned Safeguard
The withdrawn executive order would have established a voluntary preview process, allowing federal officials to review AI systems before their public deployment. While characterized as a light-touch approach that preserved industry autonomy, the measure represented one of the few concrete steps toward ensuring that rapid AI development included basic national security vetting.
The proposal did not impose mandatory restrictions or impose significant compliance burdens—it merely asked companies to show their work to relevant government agencies before release. Even this modest guardrail proved unacceptable to tech industry leaders.
Silicon Valley's Coordinated Resistance
According to sources, AI executives, with David Sacks described as leading the charge, organized what amounted to a revolt against the executive order. Their opposition was sufficiently forceful that President Trump ultimately chose to abandon his own policy initiative rather than proceed against industry wishes.
The decision highlights a structural imbalance in technology governance: well-resourced industry actors have direct access to policymakers and can mobilize quickly to block regulations, while public interest advocates, workers, and communities affected by AI deployment lack equivalent influence.
Decision-Making and Industry Access
Sources also reported that President Trump contacted Mark Zuckerberg following the decision, further illustrating the direct channel between tech leadership and executive decision-making. The episode reveals how policy outcomes in AI governance may depend less on democratic deliberation or expert consensus than on which industry figures have the president's ear at critical moments.
This pattern—where the "last person in the room" shapes major policy decisions—raises questions about whether technology regulation can serve public interest goals when industry has such disproportionate influence over the decision-making process.
Why This Matters:
The collapse of even a voluntary AI safety review process demonstrates how concentrated power in the technology sector can override government attempts at oversight. When industry opposition alone can kill a light-touch regulatory proposal, it suggests that meaningful democratic control over transformative technologies faces structural obstacles. The decision leaves no formal mechanism for national security review of advanced AI systems before public deployment. For workers, communities, and citizens who will live with the consequences of rapidly deployed AI systems—from labor displacement to algorithmic discrimination—this outcome means decisions about powerful technologies will be made entirely by private companies with no requirement for public interest consideration. The episode also illustrates how access and influence, rather than expertise or democratic process, may determine technology policy outcomes.