Anthropic and its two closest peers, OpenAI and Alphabet, are accustomed to rolling out upgraded and more capable models every few months. For now that’s stopped. Here things stand without some coherent government leadership. The industry’s investment-heavy business approach can quickly blow up if it can’t release its products to customers.
Yann LeCun, Meta’s former chief AI scientist, is among a minority who say America’s AI leaders only hoist themselves by exaggerating both the progress and danger of their latest models. Unfortunately, government fears don’t have to be accurate or well-calibrated if they cut to the heart of government’s prestige and legitimacy. The large potential benefits to society of AI Washington might willingly throw away to avoid such risks to itself.
People have been identifying and failing to identify objects in the sky forever. This has made the UFO miasma a ready resort, since early in the last Cold War, for the U.S. government to misdirect public attention from its own miscues. So far 100% of identified objects haven’t been alien spacecraft, yet people like Steven Spielberg treat the simple fact of unidentified sightings as proof not only of alien visitation, but of government conspiracies.
Americans are especially prone to such invented suspicions when they sense that sources they should be able to trust are lying to them. This problem our politicians and press have recently amplified—lies by Donald Trump’s enemies about his Russian connections and Hunter Biden’s laptop, lies by Mr. Trump about the 2020 election, lies about Covid, about the value and shortcomings of vaccines, about Jeffrey Epstein, about President Biden’s cognitive status, about UFOs.
Even allowing that some government purposes require lying, the natural course of AI adoption ought to improve matters. Sadly, our current class of disinformation-addicted bureaucrats and politicians can also use AI to make things worse. Here it may be useful to witness the still-lacking progress in domestic drone defense. It’s not too soon for Americans to start thinking about how to elect a president who can talk sanely and confidently about how we’re going to survive this future.
Also warning against strict government regulation of AI is the Editorial Board of the Washington Post. Two slices:
Perfect security is a dangerous delusion. That’s something close to a gospel truth for cybersecurity experts, who work from the understanding that their work is never finished. This appears not to be the mindset at the highest reaches of the U.S. government, which is trying to regulate away dangers that cannot be regulated away.
…..
Gatekeeping advanced models only entrenches a false sense of security. Competitors to Sol and Fable trail by months, if that. Chinese open models are closing the distance seemingly every week, and Beijing probably has access to models Americans haven’t seen.
The vulnerability found in Fable, Anthropic correctly noted, already existed in rival models the government has left untouched. But the conclusion the Trump administration apparently reached — to apply the same controls to all frontier models, including GPT-5.6 — is the wrong strategy.
When more than 150 cybersecurity professionals signed an open letter demanding the controls on Anthropic be lifted, their point was not that the danger is imaginary. It’s that stripping the best tools from the good guys while adversaries race ahead anyway leaves the country less safe.
