Zuckerberg admits he used employee data to train AI before firing them in mass

Mark Zuckerberg confirmed that Meta used intelligence data, skills, and employee interactions to train artificial intelligence models while they were awaiting mass layoffs. The episode, internally nicknamed '28 days of hell', affected approximately 8,000 workers and exposes the ethical limits in data collection by large technology companies.

How did Meta extract employee data to train AI?

According to reports of former employees obtained by Kotaku, the company implemented a 28-day transition period between the announcement of layoffs and effective exit. During this interval, employees were instructed to continue working normally, participating in meetings, code revisions and creative processes — all while Meta recorded each action to feed their machine learning algorithms.

Practice raises a warning similar to that discussed in IA evaluates its performance? Your name may cost promotion, showing how AI systems can use personal data without explicit consent.

What was the '28 days of hell' before layoffs?

Officials described the period as 'psychologically distressing'. Knowing they would be fired, many tried to prove their worth, but the Meta had already decided the cut. Contributions in these days were converted into training data for internal AI models, including CodeGen code assistant and recommendation systems. The company did not inform employees about this collection, which is a possible violation of privacy according to digital law experts.

Why is using employee data for AI worrying?

In addition to the lack of transparency, the case exposes an imbalance of power: employees who depend on wages can be coerced to provide data without knowing their fate. Meta is not the only one — giants like Google and Amazon have already been accused of using worker data to train AI, but Zuckerberg’s case is emblematic for occurring exactly during layoffs. Experts point out that this can open precedents for labor processes and regulatory fines, especially in Europe under GDPR.

How can workers protect their AI data?

Experts recommend that technology professionals review work contracts and company data policies. Many agreements contain vague clauses on 'use of data for product improvement', which allow practices like this. Industry trade union organizations require clearer laws requiring informed consent to collect performance data — especially when there are ongoing layoffs.

Frequently Asked Questions about employee data in AI

Did Meta use employee data legally fired?

Legality depends on the jurisdiction and the terms of consent signed by the officials. In the USA, where the Meta is thirsty, practice is questionable but not explicitly prohibited; in Europe, it could violate GDPR for lack of transparency.

Did the data include sensitive personal information?

According to reports, the data included productivity metrics, interactions in meetings and excerpts of code, but not medical or financial data. Still, experts consider that any unauthorized collection of performance data is ethical and legally problematic.

How do I know if my employer collects my data to train AI?

Check the privacy policy and employment contract — look for 'use data', 'machine learning' or 'product improvement' clauses. If there is no explicit mention, question HR. Tools like data auditing can also help.