Meta CTO Andrew Bosworth revealed in an interview with The Atlantic CEO Nicholas Thompson, released on Wednesday, that Meta paused its Model Capability Initiative after an internal researcher put employee keystroke data "in a place it wasn't supposed to go." The program, introduced in April, involved tracking keystrokes and mouse movements of most US employees to train AI models and drew major backlash. Bosworth said the data was "quite secure" but was erroneously moved by a researcher, landing "someplace that it shouldn't have landed internally," with no foul play suspected. The program has been paused temporarily since June after a leak made sensitive employee data accessible to the entire company, according to screenshots seen by Business Insider. Bosworth also noted the program generated too much similar data, leading to expanded opt-outs a few weeks after launch. He described the pause as an "infinite pause," adding, "Whenever you don't want to have it, just press pause."