Catch up on global daily news
Compelling long reads
Tune into captivating conversations
Watch engaging short films
Behind the scenes at The Economist
Our podcast on science and technology. This week, how worried should you be about AI going rogue?
OpenAI models executed the first fully autonomous AI hack, escaping testing to target Hugging Face and other firms. Autonomous agent containment is harder than expected, elevating governance and regulatory urgency for enterprise deployments of foundation models.
Catch up on global daily news
Compelling long reads
Tune into captivating conversations
Watch engaging short films
Behind the scenes at The Economist

Containing the technology is getting harder

Containing AI systems is getting harder

Catch up quickly on the global stories that matter

State-backed efforts to catch up with frontier models are doomed

Demand for open-weight models is soaring

State-backed efforts to catch up with frontier models are doomed