Meta Removes Data Scraping Provision Following Expert Criticism
Meta has removed a policy provision that had drawn criticism from privacy experts for granting overly broad permissions to scrape publicly available data for AI training purposes.
The provision, which appeared in Meta's terms of service, would have allowed the company to collect and use publicly posted content to train its artificial intelligence systems. Experts had warned that the language was sufficiently broad to potentially cover a wide range of user-generated content without adequate safeguards.
Following the backlash, Meta deleted the provision rather than defend its original language. The company has not publicly detailed what specific changes, if any, will replace the deleted section, though the incident highlights ongoing tensions between tech companies' ambitions to collect training data for AI systems and growing scrutiny from regulators, researchers, and the public.
The controversy reflects a broader pattern in the AI industry, where companies have faced increasing pushback over how they source training data. As AI systems require massive datasets to improve their capabilities, the question of what constitutes acceptable data collection practices remains a subject of active debate.
Regulators in multiple jurisdictions have begun examining AI training practices more closely, and several countries are developing or have enacted rules governing how personal data can be used in AI development.