You're tuned in to Tech Beat, let's get into it.
Anthropic is taking a notable step toward AI transparency, unveiling a text watermarking system designed to comply with the EU AI Act. The scheme works by subtly steering Claude toward certain inconsequential word choices — ones that, taken together, form a detectable signature. It's a quiet but meaningful shift in how AI accountability might actually work in practice.
Staying with AI, researchers are reporting genuine progress on the Riemann Hypothesis, one of mathematics' most stubborn unsolved problems, with AI tools from companies including Anthropic and OpenAI playing a supporting role. The hypothesis has stood unchallenged for over a hundred and sixty years, so "closer" is doing real work in that headline — but the direction of travel is worth watching.
And out of France, the country's top court has struck down a law that would have banned children from social media entirely. The ruling, handed down in two thousand twenty six, doesn't end the debate — it redirects it. Legislators will now have to find a more surgically precise approach to protecting minors online, a challenge that has stumped governments across the democratic world.
Keep surfing. Tech Beat out.
