The State Of AI or "On How Not To Do AI"

Woo, AI so cool! Nope, here is a collection of discriminating/abusive/erroneous/evil AI applications.

[thestateofai]

Why Meta’s latest large language model survived only three days online

Galactica was supposed to help scientists. Instead, it mindlessly spat out biased and incorrect nonsense.

Interesting Twitter thread with a few examples and their pitfalls: https://twitter.com/Michael_J_Black/status/1593133722316189696

[thestateofai]

Troubling Study Says Artificial Intelligence Can Predict Who Will Be Criminals Based on Facial Features

A new study says computers can tell whether you will be a criminal based on nothing more than your facial features.

Using AI to predict criminal behavior from facial features. Welcome back to the Medieval age. Big NONO.

[thestateofai]

In the AI era, your voice could give away your face

MIT researchers trained a machine learning model to reconstruct a very rough likeness of someone’s face based only on a short audio clip.

Using AI to “reconstruct” (read, imagine) faces from voice only. Big NONO.

If you want more insights on why this application has dangerous/discriminating implications take a look at this Twitter thread.

[thestateofai]

This startup is using AI to give workers a “productivity score”

In the last few months, millions of people around the world stopped going into offices and started doing their jobs from home. These workers may be out of sight of managers, but they are not out of mind. The upheaval has been accompanied by a reported spike in the use of surveillance software that lets…

Using AI to monitor employees’ productivity at work. Big NONO.