"Algorithmic microaggressions": It's a thing. Here's why I think it's a good term and some reflections. The attached video is an example.
➤ What is a microaggression?
👉On one definition, microaggressions are subtle acts of exclusion (link below). They can be hostile or innocent. Either way, they can make their targets feel that they don't belong, inferior, invisible, and so on.
👉Microaggressions may seem small and insignificant, but they accumulate. They can decrease confidence and even lead to anxiety, depression, and trauma. It's death by a thousand paper cuts.
➤ Examples from Wikipedia
☢️"All Asian-Americans look alike"
☢️A Korean man asking for a drink in a bar being ignored by the bartender, or the bartender choosing to serve a white man before serving the Korean man.
☢️Asian-Americans are made invisible when US race discussions focus exclusively on white/black issues
➤ Algorithmic microaggression
🤔Look at the video below. Soap is dispensed when white hands are under the sensor. Soap is not dispensed when black hands are under the sensor.
🤔The black hands are literally invisible to the sensor.
🤔To me, that seems very analogous to good old microaggressions.
➤ The risk: Microaggression at scale
🔍The example in the video is unsurprising
🔍AI algorithms typically don't work as well for minorities. Minorities are often not seen by algorithms, not recognized as human, don't get the same quality of service, etc.
🔍But AI algorithms are also ubiquitous. They are already everywhere and we are just getting started.
🔍The risk is microaggressions at scale. Invisibly, inferiority, lack of belonging everywhere you go and even in your own home. The psychological impacts could be catastrophic.
💪Yet another great reason to do AI responsibly!
➤ Who is at fault?
❌Not the algorithm itself. Let's avoid anthropomorphizing and responsibility shifting.
✔The responsibly for algorithmic microaggression lies with those who stand behind the technology.
✔It includes the companies and people that develop the technology. It also includes those who buy them, invest in them, insure them, or otherwise support them financially.
✔Responsible development, due diligence, and procurement of AI are crucial.
➤ Join the conversation about this post on LinkedIn, here
Links and references: