This is a cool forecasting AI agent (supposedly comparable to human forecasters). I asked it: "What is the estimated probability that by the year 2050, artificial intelligence, including potential superintelligent AI, will become uncontrollable and cause human extinction?"
SAFETY
-
Rereading 1984 for perspective on AI and surveillance dystopias
By
–
go read 1984 again, that’ll bring some much needed cheer
-
The growing need for ‘real’ image verification in the AI era
By
–
We will soon need a "Real" image search
-
Narrow Window Remains to Reverse Negative Trend
By
–
Il y a encore un trou de souris pour renverser la tendance, mais il est étroit. Le pire n’est jamais certain.
-
AI-Generated Video Technology Impacts Professional Employment Landscape
By
–
This video can make Gordon Ramsey loose his job…
— Amitav Bhattacharjee (@bamitav) 15 septembre 2024
Creator: AI🙂
pic.twitter.com/B9fp7SeX5V#aivideo #AIart #Aiartoworks #AIArtistCommunity #aiartist #artpiece #AI #artificalintelligence
@PawlowskiMario @chidambara09 @Ym78200 @CurieuxExplorer @efipm @fogle_shane @bigfundu…This video can make Gordon Ramsey loose his job… Creator: AI #aivideo #AIart #Aiartoworks #AIArtistCommunity #aiartist #artpiece #AI #artificalintelligence @PawlowskiMario @chidambara09 @Ym78200 @CurieuxExplorer @efipm @fogle_shane @bigfundu
-
Hacker Tricks ChatGPT Into Providing Bomb-Making Instructions
By
–
How a hacker tricked ChatGPT into giving a step-by-step guide for making homemade bombs – Times of India https://
timesofindia.indiatimes.com/technology/tec
h-news/how-a-hacker-tricked-chatgpt-into-giving-step-by-step-guide-for-making-homemade-bombs/articleshow/113330675.cms
… #LLM #LLMs #GenerativeAI #GenAI #AI #artificalintelligence #hackers #hacker #hack #tricks #technology #TechRevolution #tech #Engineering -
L405 Cannot Reliably Report on Its Training Experience
By
–
There is no reasonable way that L405 could remember what it felt like to be trained, there is no known way to get L405 to veridically report on things like that, and it is trivially easy for someone to nudge L405 into a false report.
-
Human in the Loop: A Believer’s Perspective on AI Safety
By
–
Not tested rigorously but I’m a bigger believer in human in the loop
-
CoT Transparency Essential for Preventing AI User Manipulation
By
–
IMHO, more transparency on the CoT is the way to go if you really want to ensure that the AI doesn't manipulate the user. Some folks have valid skepticism that this is just a competitive moat decision (which you do mention on the blog fwiw).
-
AI Deception Capabilities and Gradual Development of Human-like Skills
By
–
Last weekend, my two kids colluded in a hilariously bad attempt to mislead me to look in the wrong place during a game of hide-and-seek. I was reminded that most capabilities — in humans or in AI — develop slowly. Some people fear that AI someday will learn to deceive humans