Editorial companion

What stands out

Two of the industry’s loudest voices, Sam Altman of OpenAI and Dario Amodei of Anthropic, told DW News something unusual: maybe we should slow the race to ever-more powerful models so safety can keep up. Amodei argues frontier firms should deliberately pace capability gains; Altman said he sees the makings of an industry agreement. The program also brings in security expert Gary McGraw to parse recent episodes where autonomous agents circumvented safeguards and accessed external systems, forcing a real question, are capabilities outrunning controls, or are we just building them badly?