Last week researchers from Stanford University released a study for which they had created an algorithm based on deep neural networks, which is able to determine whether an individual is heterosexual or gay based on just one picture. The success rate of this algorithm for men is 81% and 71% for women. The use of similar systems by governments or companies to detect facial features is becoming more and more common and this research highlights the threat to the privacy and potentially the safety of gay men and women. By showing the danger beforehand the researchers hope to protect those individuals (Wang & Kosinski, 2017).
However good their intentions may have been, not everyone is happy with their work. The journal that was going to publish the paper has placed it under ethical review and the researchers themselves have received death threats. This showcases a bigger problem related to AI and ethics since the use of AI by social scientists is becoming more widespread but good ethical guidelines are lacking and can be decades old. Other studies have also gotten into trouble, like NamePrism. This app can determine ethnicity and nationality based on just a name with the intended use of preventing racism, but it can also be used for the exact opposite (Chen, 2017).
Some progress is now being made towards creating ethical standards, which includes the researchers behind NamePrism who have teamed up with others to create Pervade. They aim to put together a clear and transparent process for AI research (Chan, 2017). OpenAI is another non-profit aiming to create AI that benefits society as a whole and warn them about the (ethical) downsides of AI development. Co-founder Elon Musk has even called the progressive development of AI the biggest existential threat to humanity. Luckily we are not there yet, but creating some ethical guidelines and rules concerning AI research and development seems like a good idea for the sake of all of us (Dowd, 2017).
Sources:
https://www.wired.com/story/ai-research-is-in-desperate-need-of-an-ethical-watchdog/
https://www.vanityfair.com/news/2017/03/elon-musk-billion-dollar-crusade-to-stop-ai-space-x
