A data donation experiment by AlgorithmWatch shows that Welt, a brand of right-wing news conglomerate Axel Springer, largely dominates the official news section.
Read more
Using thousands of data donations from AlgorithmWatch’s Instagram monitoring browser plug-in, German newspaper Süddeutsche Zeitung showed that posts from the far-right appear higher on users’ timelines.
Read more
A feature on LinkedIn automatically rates candidates applying from another EU country as “not a fit”, which may be illegal. I asked 6 national and European agencies about the issue. None seemed interested in enforcing the law.
Read more
Google, Facebook and other very large tech companies built a lobbying system that, under the guise of furthering research or journalism, binds intellectuals in a complex network of patronage. Transparency is long overdue.
Read moreTwitter opened its image cropping algorithm and gave prizes to people who could find biases in it. While interesting in itself, the program mostly reveals the impotence of regulators.
Read moreThe European Commission published a "Guidance on Strengthening the Code of Practice on Disinformation" aimed at large tech companies on Wednesday. The wish-list of measures forgoes any enforcement mechanism.
Read moreAutomated systems from Apple and Google label characters with dark skins “Animals”.
Read moreIn Europe, several automated systems, either planned or operational, actively contribute to entrenching racism. But European civil society literally lacks the words to address the issue.
Read moreAutomated translation services tend to erase women or reduce them to stereotypes. Simply tweaking the training data or the models is not enough to make translations fair.
Read moreAn experiment by AlgorithmWatch, Groene Amsterdammer, Pointer and NOS in the Netherlands shows that Instagram probably prioritizes images of faces and pushes down images that contain text.
Read moreRouting services from TomTom and Google Maps encourage car drivers to use cycle streets, which is illegal. Both companies said they would update their databases, but this is unlikely to completely solve the problem.
Read more
Groups of teenagers use weaknesses in the notification systems of Facebook and Instagram to take over accounts and harass users. The upcoming DSA regulation addresses the issue but is likely to fall short.
Read more
A review of 338 AI-powered medical devices approved in Europe and in the United States reveals holes in the European review process.
Read moreAmid the Covid pandemic, governments and corporations stepped up the deployments of automated systems. Civil society initiatives attempted to keep some of them in check.
Read moreThe Nutri-Score summarizes basic nutritional information on a 5-letter scale. Despite its many qualities, it faces a strong backlash that could hold a lesson for operators of automated systems.
Read moreIn a seemingly routine case at the Amsterdam court of appeal, a judge ruled that it was acceptable for a municipality to use a black-box algorithm, as long as the results were unsurprising.
Read moreAn experiment reveals that Microsoft Outlook marks messages as spam on the basis of a single word, such as “Nigeria”. Spam filters are largely unaudited and could discriminate unfairly.
Read moreAn experiment by AlgorithmWatch shows that online platforms optimize ad delivery in discriminatory ways. Advertisers who use them could be breaking the law.
Read moreAn experiment shows that Google Translate systematically changes the gender of translations when they do not fit with stereotypes. It is all because of English, Google says
Read moreOnline platforms often provide data that is riddled with errors. Rather than launching quixotic attempts at fixing them, researchers increasingly investigate why platforms bias their data.
Read moreWe usually do not write about newly-released software, especially when there is no way to audit it. But the hype over GPT-3, a natural language generator, was such that several readers asked for a review.
Read moreAs part of our #LeftOnRead campaign, several researchers testified to the reluctance of online platforms to provide useful data. Many resort to studying Twitter, which is more accommodating than most.
Read moreUdbetaling Danmark was created in 2012 to streamline the payment of welfare benefits. Its fraud control algorithms can access the personal data of millions of citizens, not all of whom receive welfare payments.
Read more
