MSN's AI editor confused a mixed-race member of the band Little Mix with another and staff were told to delete articles about the mix-up should the AI find them
Firm's plan to replace editors with AI backfires after wrong image of musician is published — Microsoft's decision …
Little Mix members bear the direct harm — a public misidentification along racial lines on a major news portal — while the remaining MSN staff are put in the position of scrubbing coverage of their own employer's mistake.
Second-order effects
Publishers like the Guardian, whose journalism feeds Microsoft's curation surfaces, now have concrete evidence that automation errors attach themselves to their brands, strengthening their hand when negotiating placement and accountability terms with Microsoft Start.
Third-order effects
If the pattern holds across the Tay chatbot, the Start poll, and this misidentification, AI curation at scale becomes untenable without human-in-the-loop review for identity-sensitive content — and platform liability for automated editorial errors moves onto regulators' agendas.
The trend: Microsoft keeps trading editorial judgment for AI-driven curation scale, and each public failure — from its manipulated chatbot to the Start poll to this misidentification — raises the real cost of that trade.
The humans supervising the racist robots have been told the racist robots may try and write stories about how they're racist, and if they do, the humans are to intervene to stop it https://www.theguardian.com/ ... https://twitter.com/...
Last week we revealed Microsoft is replacing its human news editors with artificial intelligence software. A week later the robot journalist is facing accusations of racism after it confused different mixed-race members of Little Mix. https://www.theguardian.com/ ...
Last month @Microsoft fired hundreds of journalists in the middle of a pandemic & fully replace them with an AI software. It backfired as their AI cannot tell mixed-race individuals apart. 🤦🏻♀ ️ https://www.theguardian.com/ ...
“Because they are unable to stop the new robot editor selecting stories from external news sites, human staff have been told to stay alert and delete a version of this article if the robot decides it is of interest and automatically publishes it” https://www.theguardian.com/ ...
This is why AI obsession sucks. If humans are struggling with the nuances of race & language, why would robots be able to get it right? The people programming these algorithms come from non-diverse backgrounds and it's gonna SHOW itself in so many stupid ways https://twitter.com/…
For journalists who thought that pissing off woke colleagues was their biggest worry, meet your new robot overlords: “Even if they delete it, the robot editor may overrule them and attempt to publish it again.” https://www.theguardian.com/ ...
‘The remaining human staff have been told to stay alert and delete a version of this article if the robot decides it is of interest and publishes it ... even if they delete it, the robot editor may overrule them and attempt to publish it again’ LOL https://www.theguardian.com/ ..…
MSN uses algorithm to select news articles; editorial algo has racist bias; The Guardian reports it; Microsoft's response is to try to censor their reporting: Microsoft then continues with plan to fire human staff, replace with racist AI. Nicely played. https://www.theguardian.co…
“Staff have already had to delete coverage criticising MSN for running the story about Little Mix with the wrong image after the AI software decided stories about the incident would interest MSN readers” https://twitter.com/...
You can try to explain what satire is, or you can just wait for actual events and point at them and say “if a fiction writer had come up with this, say.” https://twitter.com/...
Staff at MSN have also been told to await the publication of this Guardian article and try to manually delete it from the website, because there is a high risk the Microsoft robot editor taking their jobs will decide it is of interest to MSN readers. https://www.theguardian.com/ …
The thing about AI is it just reproduces the biases, prejudices and structures of power within the society it emerges from and I wish this was at the forefront of conversations about AI - along with an acceptance that technology is *not* amoral. https://twitter.com/...
I think we really need a better word than robot here. Robot assigns too much agency and capability. Excel formula? Because then it'd be clear there's no such thing as a racist excel formula, just a racist person who writes formulas. https://twitter.com/...
Newspaper is trolling newspaper with racist robot by publishing article about how racist robot is publishing articles about racist robots in the hopes that racist robot will republish racist robot article. https://twitter.com/...