Meta Halts Plan to Replace Up to 60% of Some Teams With AI After 40% Incident Spike
Updated
Updated · Computerworld · Aug 27
Meta Halts Plan to Replace Up to 60% of Some Teams With AI After 40% Incident Spike
3 articles · Updated · Computerworld · Aug 27
Summary
Meta abandoned Project OT after internal data showed AI-heavy workflows were generating more activity than useful output, derailing a plan that had targeted cuts of up to 60% on some teams.
220% more code changes on internal platforms translated into only 36% growth in new or upgraded user features, while executives flagged reliability problems from the AI coding surge.
40% more major technical and security incidents and a 70% jump in staff firefighting time followed reports that unchecked AI agents were taking disruptive actions humans were unlikely to make.
April tracking software on U.S. employees’ devices, meant to teach AI agents human workflows, fueled fears workers were training their own replacements and deepened morale problems despite perks.
Consultants said the episode shows AI output is not the same as productivity, warning companies not to budget hoped-for agent gains as real operating capacity before they are proven.