Microsoft has come under mounting pressure to refrain from using Gender Labels for its artificial intelligence (AI) and facial recognition technologies. However, Microsoft is not the only company facing pressure to stop using Gender Labels. Many other major tech companies like Amazon and IBM are also subject to scrutiny with respect to their usage of Gender Labels for their respective AI and facial recognition services.
Following Google’s steps, Microsoft under pressure to stop using Gender Labels in AI
What is Gender Labels?
In the context of artificial intelligence (AI) and facial recognition, the usage of Gender Labels is considered controversial and discriminatory. There are certain AI tools that can classify the person as either a “man” or a “woman,” solely judging by the person’s face, which could lead to unfair gender bias.
AI algorithms trained using faulty or intentionally-biased datasets can misclassify people’s gender. One study successfully connects the dots between AI and sexism:
Researchers trained AI using cooking-related images. In these images, women were 33% more likely to appear. The study found that AI was 68% more likely to predict a woman was cooking, even though that wasn’t actually the case.
Google swings into action
Last month, Google suspended its usage of gender labels in its image recognition AI. Google shared an update that its AI algorithm will now classify every individual as a “person.” Google will introduce these changes to Google Cloud’s Vision API:
“As you know, the cloud Vision API can perform feature detection on a local image file for the purpose of identifying persons by sending the contents of the image file through ‘LABEL_DETCTION’.”
“Currently, when you request the API to annotate an image with labels, if you use this feature on images with people, it may return labels describing them in an image with gendered terms, like ‘man’ or ‘woman’.”
Google acknowledged that a person’s gender cannot be judged by appearance which is why they have decided to remove Gender Labels to align with Google’s AI Principles. Thus, the onus is on Microsoft, Amazon, IBM, and other technology companies to follow in the footsteps of Google and stop using Gender Labels in AI and facial recognition services.
