A password will be e-mailed to you.

Facebook AI misidentifies black men as ‘Primates’, issues apology

Facebook

Source: Al Jazeera

Facebook is one of the world’s largest technology companies in the world and as we know, it has previously been into several controversies because of its Artificial Intelligence systems. Well, Facebook’s Artificial Intelligence system screwed the company yet again by labeling black men as “primates”.

Yes, you read that right, Facebook’s AI misidentified a video of black men as primate content and this is something highly offensive to the black community in its entirety. According to recent reports, users who recently watched a video of black men posted by Daily Mail on Facebook were prompted by the system asking them if they’d like to keep seeing videos about Primates, as noted in a report by Engadget. Well, this is something very serious for the black community and the social media conglomerate has also apologized for the same, calling it an ‘unacceptable error’ as mentioned in a statement issued to the publication.

According to a report by Engadget, the technology giant has already disabled its Artificial Intelligence recommendation feature followed by the incident that was responsible for the prompt. This action was taken as a precautionary measure to avoid such incidents from happening in the future. The company apologizes to everyone who saw the offensive recommendation.

Incidents like this are proof that Artificial Intelligence on social media isn’t as accurate as we think it is, but after multiple incidents with Twitter and Google, and Facebook itself, these companies are forced to believe that their Artificial Intelligence systems are far from perfect and there is a lot of work still left to do.

For instance, in a previous incident, Google mistakenly labeled photos of Black men as “gorillas” and as a solution, the technology giant censored the word “gorillas” from its AI system which is not a very reliable fix.

 

Twitter’s image cropping mechanisms were found to be biased against Black men and were only highlighting White men. The system was faulty to the extent that it was also preferring beautiful women to display on the platform as an automated image cropping mechanism.

So, Facebook’s mistake is huge, but the fact that it agrees that its AI system is not yet perfect and that they apologized and disabled the feature effective immediately shows that this was an honest AI mistake.

The company has said to improve on its Artificial Intelligence systems and they have also hired professionals to properly label people with different skin tones, so that the AI systems can learn what people with different ethnicities look like under various lighting conditions, as mentioned in a report by Engadget. Thus, Artificial Intelligence is still a long way to go!

 

Comments

comments

No more articles
Send this to a friend