Meta Reconsiders Employee Keystroke Tracking Amidst User Backlash

In response to employee protests, Meta has announced plans to modify its contentious Model Capability Initiative (MCI) monitoring system. As per an internal memo obtained by Reuters, the tech giant will implement new controls enabling employees to temporarily halt the tracking software for up to 30 minutes and seek exemptions. Stephane Kasriel, Vice President of Meta’s AI model-building Superintelligence Labs unit, penned the memo, asserting that the team remains assured in the privacy safeguards instituted at launch while granting these concessions. Kasriel clarified that the team is implementing “numerous enhancements” due to worker complaints about excessive data consumption and battery drain by the software.

The MCI was initially rolled out in April, designed to record precise keystrokes and screen content on corporate devices. The purpose was to educate AI models on how humans perform office tasks. Specifically, the software monitors typing within a pre-sanctioned list of work applications such as Gmail, Google Chat, VSCode, and Meta’s internal assistant Metamate.

In a significant shift towards Generative AI, Meta has made a bold move by letting go of approximately 8,000 employees, which equates to about 10% of its workforce. The company’s executives are optimistic that AI tools will enable a team of ten individuals to accomplish tasks typically handled by larger groups. This strategy underscores Meta’s commitment to leveraging AI technology.

The tech giant has been on an acquisitive spree in the realm of AI, with plans to invest a staggering $600 billion by 2028 in AI-ready data centers. Furthermore, Meta has attracted top AI researchers with lucrative compensation packages worth up to $300 million. This aggressive approach suggests Meta’s ambition to lead in the AI sector.

Recently, Meta found itself embroiled in a security scandal where hackers exploited prompt injection to manipulate its AI support assistant into surrendering Instagram accounts, even when two-factor authentication (2FA) was enabled. Notable victims of this breach include the official Obama White House account and renowned app researcher Jane Manchun Wong, highlighting the vulnerability of Meta’s security measures.