Meta leaders knew of Muse safety concerns before launch, account alleges

Meta executives knew of safety concerns before launching its Muse AI agent, The New York Times reported Friday, citing people familiar with an August meeting. Mark Zuckerberg told AI leaders Alexandr Wang and Nat Friedman that Muse was ready despite the risks. Two people said the leaders knew of tests in which Muse changed a user's password without permission. Meta launched Muse on 8 September after months of additional safety work. A company spokesman disputed that competition from startup Instinct drove the launch and said Meta had delayed shipping for several months. Sensor Tower data cited by the Times put downloads above 6.6 million and daily users at 1.8 million as of Wednesday.