Feds say AI favors defenders over attackers in cyberspace — so far 

U.S. cyber leaders from the FBI and DHS assess that, for now, generative AI benefits defenders more than attackers, though the advantage may be temporary; they note that state-backed actors from China, Russia, Iran, and North Korea are experimenting with LLMs with limited gains to date, while groups like China-linked Volt Typhoon employ living-off-the-land techniques and obfuscation to evade detection and complicate attribution.