The White House Put Out an AI Framework. It Doesn't Say Who's Responsible for Anything.

On March 20, 2026, the Trump administration released its National Policy Framework for Artificial Intelligence. According to Brookings Institution visiting fellows Tom Wheeler and Bill Baer, the document contains "worthy aspirations" — protecting children, promoting innovation, American leadership. Standard stuff.

But Wheeler and Baer identified the fatal flaw: the framework never answers the most basic governance question. Who is in charge of those in charge?

The White House punts the hard decisions to Congress with a string of "Congress should" suggestions. At the same time, it says almost nothing about holding the actual AI decision-makers — a handful of tech executives accountable primarily to their shareholders — responsible for the consequences of their systems.

It's not a governance framework. It's a press release.

Brookings: We're Confusing Symptoms for Causes

Wheeler and Baer at Brookings argue the real threat isn't the sci-fi doomsday scenario of machines going rogue. It's the concentration of AI decision-making power in a tiny number of people who answer to nobody but themselves and their investors.

They propose four principles any real policy needs: accountability, access, agency, and action. The Trump framework, they say, delivers on none of them in any meaningful, enforceable way.

It's a structural critique with merit.

The UN Launched Global AI Governance. The U.S. Said No.

On September 25, 2025, the United Nations launched its Global Dialogue on AI Governance, according to the Center for Strategic and International Studies. UN Secretary General António Guterres framed it as the first time "every country will have a seat at the table of AI."

The initiative came out of the UN's High-Level Advisory Body on AI and was approved by the General Assembly in August 2025 after negotiations led by Spain and Costa Rica. Annual meetings will start at the 2026 AI for Good Global Summit in Geneva.

Most mainstream coverage didn't flag it: the United States came out in strong opposition to the entire framework at a UN Security Council debate the day before the launch, according to CSIS.

Most countries called for cooperation. The U.S. was the outlier. Whether that's principled sovereignty protection or short-sighted isolation depends on your read — but the fact remains: the world is building an AI governance structure, and America opposed it.

Meanwhile, the Guardrail-Free AI Problem Got Worse

Our previous coverage flagged the rise of open-weight models that refuse nothing. NPR's reporting from May 31, 2026 adds context: stripping guardrails used to require time and deep technical expertise. Not anymore.

Noam Schwartz, CEO of AI security firm Alice, told NPR that in recent months the process of removing safety filters from open-weight models has become "dramatically more accessible and popular." His firm conducts red-teaming and safety evaluations for model developers.

These aren't fringe tools. Open-weight models are now being released by OpenAI, Alibaba, and China's DeepSeek. They have capabilities close to the best proprietary models. And once they're out, the original developer has zero control over what happens to them.

"Everybody can download and operate their own state-of-the-art model and use it for great things and terrible things," Schwartz told NPR.

These stripped models will answer requests on how to make explosives, plan a mass shooting, or synthesize drugs. No refusal. No flag. No record.

The White House framework mentions protecting children. It does not name a single enforcement mechanism for this specific, documented threat.

Yale Is Workshopping AI-Weighted Voting. Someone Should Know That.

At the fourth International Workshop on Reimagining Democracy hosted at Yale's Institution for Social and Policy Studies in January 2026, attendees debated ideas including algorithmic systems that would assign different weights to different citizens' votes based on voting history and "consistency."

Science fiction author Eugene Fischer pitched the concept as "loyalty points for democracy" — reward people for persistence or thoughtfulness, give their votes more weight over time.

Harvard's Bruce Schneier, who co-organized the event, told Yale's ISPS that this was "the most AI-heavy year" the conference had seen — "not by design, but because that's what's in the air."

The idea that an algorithm should determine how much your vote counts is advancing in academic circles with minimal public scrutiny.

What Mainstream Media Is Getting Wrong

Left-leaning outlets like the NYT are framing this as a call for collective action — "we must take AI's future into our own hands" — without naming who "we" is or what specific power structures need to change.

Fox News is running Pope Leo XIV's AI encyclical warnings about souls, which addresses the philosophical dimension but sidesteps the concrete policy vacuum entirely.

Neither side is spending enough time on the specific, measurable fact: the U.S. has no enforceable AI accountability framework, opposed the only multilateral effort to create one, and the most dangerous open-weight models are getting easier to weaponize by the month.

What This Means

If something goes catastrophically wrong — a mass casualty event aided by an unfiltered AI model, a financial system manipulated by unchecked algorithms, deepfake election interference — there is currently no agency, no official, and no law clearly responsible for preventing it or prosecuting it.

The Trump framework told Congress to fix it. Congress hasn't moved. The UN tried to build a global floor. The U.S. walked out.

Regular Americans had no idea any of this was happening.