The Irish Council for Civil Liberties (ICCL) did a 30-day look at real-time bidding (RTB) and found user location and activity data is exposed on average 747 times/day in the US, and 376 times/day in the EU. Even more alarming, the ICCL considers these numbers low, since Facebook and Amazon RTB broadcasts weren’t included in the count.
New report from the Future of Privacy Forum (FPF) looks at how automated-decision making (ADM) has been viewed historically in cases by national courts and by Data Protection Authorities under GDPR and previous law – and explores how this can inform what to watch for as AI and Machine Learning (ML) use grows.
The US Federal Trade Commission has announced that it is illegal under the Children’s Online Privacy and Protection Act (COPPA) for companies to deny students access to remote learning platforms if parents or schools refuse consent for student data to be used for marketing and advertising. Companies will also not be allowed to use data collected from a child with a school’s authorization for commercial purposes, and they are obligated to remove data expeditiously when it’s no longer needed.
Like trouble, bad behavior by Meta shows up whether you look for it or not. The latest is an open-source language model that was supposed to provide reliable search results because it was trained on academic papers. Alas, it was quickly withdrawn after reviewers found that it returned results that were grammatical and plausible but also incorrect, not to mention filled with “antisemitism, homophobia, and misogyny.” How can this be a surprise?