We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
Hacker News·Jun 21·5 min read
Petition against Meta's employee training data collection for ML models
TL;DR
Meta announced a program called MCI (Model Capability Initiative) to collect employee computer-use data for training AI models. Employees raised concerns about the lack of privacy reviews and vague mitigations regarding data collection.
✦ Why It Matters
Engineers should prioritize ethical data collection practices and advocate for transparency in AI training initiatives.
Key Takeaways
- Advocate for clear privacy policies in data collection.
- Ensure transparency in AI training initiatives.
- Engage in discussions about ethical data use.
- Support employee rights regarding personal data.
Full Summary
Meta's Model Capability Initiative (MCI) aims to enhance AI models by collecting detailed employee computer-use data, including mouse movements, click locations, keystrokes, and screen content. Employees expressed concerns over the absence of completed privacy reviews, which are necessary for processing personal data, and noted that the proposed privacy mitigations were unclear.
Leadership's limited confidence in these measures was highlighted by the selective opt-out option available only to executives. This situation has led to a petition demanding that Meta halt the collection of such data, emphasizing the need for ethical standards in AI development.
The initiative raises significant questions about employee privacy and the ethical implications of using personal data for machine learning. As a result, there is a growing call for transparency and accountability in how companies handle employee data.