My Little Pony and racism – wait what? [The Atlantic]
As unbelievable as it may seem to most of us, the famous cartoon of magical horses has a strange online following which is now boiling to the surface. Since 2010, a certain adult fan-base of the My Little Pony has been using the online platforms to share messages of white supremacy and racism. Some of the other fans chose to look the other way, but with the Black Lives Matter movement gaining more momentum recently, real fans have decided enough is enough. Support messages for BLM have been popping up and racist messages to stir violence removed. So even if you live in an online fantasy world, reality will catch up with you to take responsibility for your actions. No excuses anymore.
Facial recognition isn’t picture perfect [The New York Times]
A man in Detroit got arrested by police without any explanation or context. Only once the interrogation started a day after his arrest, both him and the police realised the algorithm linking Mr Williams’ face to a robbery was completely wrong. False positives in facial recognition software are not as rare as you think and happen more frequently with Black than white men. A biased and imperfect system should not be an absolute tool, and Mr Williams is still waiting for an apology…
Real life Minority Report profiling [The Verge]
In the vein of sketchy algorithms – but with less action than Tom Cruise in Minority Report – AI experts are calling attention to algorithms that claim to predict crimes based on data like facial scans and criminal statistics. The Coalition for Critical Technology has published an open letter that criticises such work as “scientifically illiterate” and perpetuating prejudices and racism. The letter is in response to news that Springer, the world’s largest publisher of academic books, is planning on publishing such a study that claims to be 80% accurate with no racial bias. Call us sceptical.
AI for the assist [Wired]
On to more positive uses of AI and machine learning – Acronis, a data storage, backup, and cyber-protection company – is working on using its footage of football games to help premier clubs around Europe play their best. Football in Europe is a bit late to the game when it comes to data-led performance reviews. Acronis has stepped in to use footage for machine-learning algorithms to spot patterns which can help managers and coaches make important game decisions quicker. The algorithm can also be used off the field to see when audience members are most engaged to better determine when to show adverts. Some football fans will be pleased to know that the company is also working to use the software to pick new talent and to help rehab injured players.
In case you haven’t had enough…
Another Tweet From Trump Gets a Label From Twitter (The New York Times)
Segway stops production, marking the end of a scooter era (The Washington Post)
The algos that steal your prose voice and language (The Financial Times)