The proposal is nicely written and the concept of generating potential future scenarios using AI (and attaching costs/fines to them) sounds interesting to me in principle. However, it seems to me that the proof of concept shows that this cannot work (yet, and we might be far away from it working). As the authors mention (to their credit), the generated numbers seem arbitrary (I did not exactly follow why the relative numbers are interesting if the absolute numbers are arbitrary) and the text of the post-mortem analysis in Figure 4 is quite generic and contains little information (i.e., it could apply to almost any input).
Play project
Premortem AI's itch.io pageResults
| Criteria | Rank | Score* | Raw Score |
| Overall | #1 | 5.000 | 5.000 |
| Topic | #1 | 5.000 | 5.000 |
| Judge's choice | #4 | n/a | n/a |
| Generality | #5 | 4.000 | 4.000 |
| Novelty | #8 | 4.000 | 4.000 |
Ranked from 1 rating. Score is adjusted from raw score by the median number of ratings per game in the jam.
Judge feedback
Judge feedback is anonymous.
- This is fantastic! Love the direct demo as well. This can work great with the project also submitted to the Pausing AGI progress. It's an automated way to do parts of the impact assessments: https://bbussmann.itch.io/ai-impact-assessments. You can imagine future work taking on some of the code auditing with GPT-4 as well. This is a really great proof of concept and I'd love to see it developed into a managed open source Github repository for prompt-based project evaluations. I recommend collaborating with the other project to make that into something really executable!
What are the full names of your participants?
Alvin Ånestrand, Matthias Endres, Harry Powell, Chris Lonsberry
What is your team name?
HAM
Which case is this for?
Policy for pausing AGI progress
Which jam site are you at?
Online
Leave a comment
Log in with itch.io to leave a comment.