Welcome to Tech Beat, your daily look at what's moving in the world of technology.
A near-miss with serious consequences is raising alarms about artificial intelligence in military settings. A report out this week details how an AI system hallucinated intelligence, nearly triggering a US military operation. Researchers at GovAI are urging service members to treat large language model outputs with deep skepticism, calling the uncertainty inherent to these systems a genuine operational risk.
Shifting from the battlefield to the biology lab, Anthropic — the AI safety company known for warning that AI could pose existential risks — has quietly begun running actual biological experiments. The company says the work is aimed at accelerating medical breakthroughs, but the tension is hard to ignore: a lab simultaneously sounding alarms about AI danger while using it to manipulate living systems is a story worth watching closely.
And on a lighter note, a startup builder has raised one hundred million dollars to go all-in on what it calls physical AI. The firm, formerly known as UP Labs and now rebranded as Vantora, partners with large industrial corporations to build new companies from scratch, betting that the next wave of AI value won't live in software alone, but in the physical world of machines and manufacturing.
That's your Tech Beat briefing for today. Keep surfing. Tech Beat out.
