Scenarios giving rise to potentially rogue AIs, from intential genocidal behavior of some humans to unintentional catastrophe arising because of AI misalignment.
Comments closedYoshua Bengio Posts
There have recently been lots of discussions about the risks of AI, whether in the short term with existing methods or in the longer term…
Comments closedI recently signed an open letter asking to slow down the development of giant AI systems more powerful than GPT-4 –those that currently pass the…
Comments closed
Co-written with my PhD student Edward J. Hu. Scaling seems to work really well. We must be cautious in pursuing research directions that build knowledge…
Comments closed(see tutorial and paper list here) I have rarely been as enthusiastic about a new research direction. We call them GFlowNets, for Generative Flow Networks.…
Comments closedWe sometimes come across catchy headlines about AI that can scare or excite us, such as “By the middle of this century, artificial intelligence will…
Comments closed
I was at the Montreal climate march on Friday Sept 24th, 2021 with a bunch of Mila folk and it got me thinking again and…
Comments closed
In my previous blog entry, I discussed how when each person or corporation maximizes their individual interests, it can easily end up with most of…
Comments closed
Many experts, but not enough people, are well aware that markets can fail to achieve the goal of maximizing long-term wealth, and indeed they do…
Comments closed
Mila’s COVI project has found itself at the centre of a public debate regarding the use of an app in the fight against COVID-19. While…
Comments closed