Responsible AI Pledge Challenge
"If anyone is feeling anxious, worried or maybe you just want a chat, please, please do not come crying to me." Sister Michael, Derry Girls
Our argument - I am a contributor to RTW - is that if we act collectively, sharing our concerns and aspirations, we have a better chance of putting AI to good use.
The problem with making pledges is that you're supposed to stick to them in perpetuity. So it was a relief when my colleague Alja and I at RTW realised that pledges, like humans, are sensitive to time. Principles are rocks which change imperceptibly, if at all, but pledges are like eddies in the surrounding waters, to be navigated until we are clear of immediate danger.
Having a moral window in which to act also shifts the emphasis from the doer to the thing being done. Acts are transitory but people, and circumstances, change.
I've been using ChatGPT a lot this year, and talking and thinking about it more. Does the provenance of an idea or work of art matter, do memories exist (or are they simply re-reruns through the model of the brain), and is AI (or not AI), a distraction we cannot afford? I've also been asking myself what the cost is to use AI, to our privacy, to people's rights, and to the environment.
I will return to these questions in later posts, but for now, here are my pledges.
With thanks to Alja, Carl, JP, Franca, Christina, Carla, Pierre, Anne and Bernard, and my former colleagues at Altruistiq with all of whom I have shared conversations about the upsides, downsides, and conundrums of recent developments in AI.