LinkedIn scraped user data for training before updating its terms of service | TechCrunch

LinkedIn has updated its terms of service to disclose that it uses U.S. user data to train generative AI models, a practice notably excluded for European users due to stricter privacy regulations. Although an opt-out toggle exists, the platform faced criticism for implementing this significant change without prior notice or a mandatory consent request. This delay undermines the principle of transparency, as users were not given adequate warning to adjust their privacy settings or leave the platform before their data was already being repurposed for AI development. The article highlights a critical issue regarding the "opt-out" model, which advocates argue is fundamentally inadequate for protecting user rights. Experts contend that relying on users to proactively monitor and disable data harvesting is unrealistic and insufficient. Instead, they demand an "opt-in" consent framework, where explicit permission is required before any personal data is utilized. This debate underscores the tension between the tech industry’s hunger for training data and the legal and ethical necessity of obtaining clear, informed consent from individuals whose information is being processed. This situation is highly relevant to open data discussions as it illustrates the complex intersection of proprietary data usage, AI training, and user sovereignty. While open data often promotes transparency and accessibility, this case demonstrates how companies can obscure data practices, using technical mechanisms like late-term updates to bypass meaningful user control. It serves as a cautionary example of how data reuse for AI can conflict with open principles if not governed by robust, upfront consent mechanisms that prioritize individual autonomy over corporate efficiency.

Source: techcrunch.com
Published on 2024-09-19