Biden's EO on AI tries to chart a middle path by letting AI development continue with modest new rules and signaling the government plans to monitor AI progress
This order extends a federal effort to guide AI development that had previously emphasized public trust, while turning that approach into agency-directed work on standards, privacy, worker protections, and safety through the administration's generative-AI order.
Its significance lies less in stopping development than in making oversight an ongoing government function. Subsequent analysis questioned whether the order's compute-based reporting trigger could reliably identify the riskiest individual models.
First-order effects
AI developers can continue advancing models while facing new federal expectations around safety, privacy, and reporting as agencies develop standards.
NIST, DHS, and other agencies become central channels for translating broad AI-policy goals into operational guidance and monitoring.
Second-order effects
Compliance and risk teams gain a larger role in model-development decisions, particularly where companies must determine whether training runs or deployments fall within emerging federal scrutiny.
A compute threshold can steer attention toward the largest training efforts, but its imperfect link to model-specific risk may leave companies and agencies debating what signals should trigger oversight.
Third-order effects
The order points toward state-mediated AI development: government oversight layered onto continued private-sector scaling rather than a blanket licensing halt.
The durability of that model remains politically contingent; later coverage of an AI action plan seeking to roll back Biden-era bias and cybersecurity measures shows how executive-branch AI rules can be revised between administrations.
The trend: AI governance is moving toward continuous, agency-led monitoring of frontier development, even as the exact compliance burden remains subject to political change.
Today's AI executive order has impressive scope. It also reads like the smashing-together of papers written by contrasting groups, one that adopts wholesale the talking points of AI firms, another that centers civil liberties. @matteo_wong & my analysis. https://www.theatlantic.c…
Biden: With AI, fraudsters can take a three second recording of your voice, I have watched one of me on a couple of occasions. I said when the hell did I sat that? [video]
Biden's AI executive order has been released in full. As I reported a few weeks ago, it requires that cloud companies tell the government if a foreigner is training an AI model that “could be used in malicious cyber-enabled activity” https://www.semafor.com/... [image]
Thank you, President Biden! I look forward to building upon on today's important executive order and continuing to drive towards bipartisan legislative action on AI. We cannot afford to wait. [video]
Biden's exec order on AI has now mandated notifications of training runs and reporting of red team safety test results for frontier models! Also, standards for bio-synthesis screening; watermarking etc. Powerful steps in the right direction. No provision for transparency though
Less spoken about U.S. Biden's Executive Order on AI is that it has a lot to unpack about privacy. That's a privacy win, even without something like #GDPR. Thread. It even uses the highly technical term “differential-privacy” (and privacy-enhancing technologies)! [image]
1/3: New AIPI polling reveals sweeping support across the political spectrum for President Biden's AI Executive Order today. An overwhelming 69% of American voters are in favor, including 64% of Republicans and 65% of Independents. Respondents were the most supportive of... [imag…
Biden's executive order on AI is a complete joke, just like the order on crypto. Why is the Biden administration so rabidly anti-innovation? Maybe if we didn't elect septuagenarians and octogenarians to Congress and President we wouldn't have to deal with this stupidity.
Just landed in Washington as part of a delegation of @EP_SingleMarket to discuss with US representatives about digital issues and our AI Act, the same day as President Biden signs the Executive Order on Artificial Intelligence. Impossible to get a better timming! 📲🇺🇸🇪🇺
More interesting AI poll results, this time related to Biden's recent executive order. Yet again, policies that are the most helpful for reducing AI X-risk (starred below, according to my personal assessment) are the most popular! [image]
Artificial Intelligence has the potential to change the way we work, learn, and create. I'm glad to see @POTUS Biden signing an executive order on AI designed to encourage innovation while avoiding some of the biggest risks. https://barackobama.medium.com/ ...
President Biden has signed an executive order that will require AI companies to “address algorithmic discrimination” and “ensure that AI advances equity.” They want to embed the principles of CRT and DEI into every aspect of AI. [image]
There's a lot to unpack in Biden's Executive Order, but it's clear that the US sees attracting global talent as key to competitiveness in AI. The UK must adopt this mindset too and here is what @StartupCltn, @ukonward and @InstituteGC propose [image]
I'm glad the Biden administration is heeding our call to include principles from the AI Bill of Rights in its new executive order. Now, we must ensure robust implementation and pass legislation in Congress to protect our fundamental rights and regulate AI systems.