Welcome to Our Research Archive
Search and filter by content type, issue area, author, and keyword
August 13, 2025
Reading the Mind of the Machine: Why GPT-5’s Chain-of-Thought Monitoring Matters for AI Safety
When OpenAI released GPT-5 last week, most coverage focused on the model’s performance gains, including fewer hallucinations, stronger multilingual abilities, and state-of-the-art results in health tasks. But buried deep in the 59-page system card is something far more consequential for AI safety: a section describing how OpenAI is monitoring GPT-5’s internal reasoning (“chain-of-thought”) in real-world…
August 13, 2025
Understanding Why the Supreme Court’s Ruling in Free Speech Coalition v. Paxton Is Narrow
With the US Supreme Court now considering in NetChoice v. Fitch whether to reinstate an injunction blocking enforcement of Mississippi’s online age-verification and parental-consent law, it’s vital to understand the narrowness of the Court’s recent ruling in Free Speech Coalition v. Paxton. The Court in Free Speech Coalition upheld, against a First Amendment challenge, a…
August 12, 2025
The Climate Beat Goes On
Last week I was contacted by two reporters at the Associated Press with a request to comment on the Department of Energy’s Climate Working Group (DOE CWG) report and the proposal by the Environmental Protection Agency (EPA) to rescind the 2009 greenhouse gas endangerment finding: We’re Seth Borenstein and Michael Phillis, reporters on the climate and environment team at…
August 12, 2025
The Climate Conversation is Changing
Later this week here at THB I’ll be publishing two important pieces — one a guest post from a climate scientist on how his work was cited in the DOE CWG report and the other exposing a major scandal in climate research. Today, I share a big pile of recommended readings. We are all lucky to have…
August 12, 2025
If the U.S. Doesn’t Set Global Tech Standards, China Will
Imagine an internet where your identity is automatically attached to everything you do—every website you visit, every click you make. That was the vision behind New IP, a proposal Chinese engineers introduced at a United Nations telecom forum in 2019. New IP would have replaced the current open internet with a government-controlled system designed for…
August 8, 2025
Is Making Platforms Responsible for Banning Australian Children’s Use a Straw Man?
Prime Minister Anthony Albanese’s Australian federal government is forging ahead with its “world-leading” legislation aimed at protecting Australian under-16s from social media harms. Last week, it was announced that YouTube—previously the beneficiary of a carve-out that excluded it from earlier versions of the legislation—will now be included in measures to “protect young Australians at a…
August 7, 2025
Securing Critical Infrastructure in the Age of AI
The Trump administration’s America’s AI Action Plan—released in July—takes a significant step toward positioning artificial intelligence as both a national asset and a defense tool. One of its most notable provisions calls for the strategic use of AI to protect America’s critical infrastructure, from pipelines and power grids to financial systems and public services. As…
August 6, 2025
Some Property Rights Archaeology
In a recent post, I discussed the interesting response to the idea that information has taken on the characteristics of common-law property: “That would be a bad idea.” Saying so has incongruence akin to looking at a rainstorm and saying it’s a bad idea. Good or bad, I believe it’s happening. And I think it’s…
August 5, 2025
Well Cited
Last week, a colleague of mine sent me a copy of an email that they had received from ClimateBrief, a UK-based advocacy journalism group. The email asked for examples of how their published research had been “falsely or misleadingly characterised” in the Department of Energy (DOE) Climate Working Group (CWG) report. That email began as follows:…
August 5, 2025
Trump’s AI Plan Is Right to Cut Red Tape. But It Must Also Confront Europe’s Data Grab.
President Trump’s new AI Action Plan gets the big picture right: America’s private sector—not government central planners—should lead AI development. But to achieve that goal, the administration will have to do more than roll back domestic red tape. It must also confront a growing threat from abroad: the European Union’s data strategy, which deliberately hobbles…