Originally published at: MIT's AI risk database exposes 700+ ways AI could ruin your life - Boing Boing
…
3 Likes
The LessWrong techbros would prefer we focus on these risks, of course. Meanwhile, the more likely results of their usual “move fast and break things” approach will end in misery for a lot of people who aren’t white or model-minority affluent cis-het males.
3 Likes
A comprehensive living database of over 700 AI risks categorized by their cause and risk domain.
please tell me that ‘A.I.’ wasn’t significantly utilized to construct this ‘comprehensive living database’ (“living”? …suppose that means one can still submit: INSERT INTO table_name (columnN, ...) VALUES (valueN, ...) as one does in ‘normal’ databases…?)
3 Likes
This… is really not true. That’s a tertiary concern, at most, for almost everyone in those circles. The kind of misery you’re talking about is the secondary concern. The primary concern is “literal human extinction or comparable suffering for everyone within the next few decades.” Which, whether you think that’s plausible or not, is definitely the right order of “Which risks are worse if they happened.”
Based on past similar efforts, I expect the reception to this database to be “It’s great that this exists, now let’s figure out how to actually solve them and hope people take it seriously and slow down the dangerous parts of the progress being made in AI.”
I (and the MIT study) are not talking about severity in and of itself but rather likelihood (and, to a lesser extent, timing).as the source of concern re: risk.
The businesspeople funding the development of “AI” have demonstrated themselves fundamentally uninterested in – if not outright happy to promote – outcomes like racial bias in algorithms or privacy violations. To that end, they’re glad for us to worry about gedankenexperiments like Roko’s Basilisk (I’m done for now, aren’t I?) as a distraction.
3 Likes
Of course, there’s a question of whether any Silicon Valley startups will actually consult this list before developing some hokey new generative AI feature to dazzle their investors with before laying off a bunch of workers for absolutely no good reason.
Of course they’ll consult it, it’s full of “move fast and break things” ideas that they’ll be having their hirelings implement as soon as possible! Hey, I bet we could use these to create The Torment Nexus!
http://www.antipope.org/charlie/blog-static/2023/11/dont-create-the-torment-nexus.html
2 Likes
That’s true, I completely agree. It’s just about the opinions of a different group of people than those you referenced in your prior comment.
Edit to add:
How dare you put all these fine unknowing BBSers in mortal peril?! /s 
Personally I enjoy the gedankenexperiments in general, and am intrigued by the concept of (for example) acausal trade in particular, but it’s hardly something I’d want people worried about.
1 Like
This topic was automatically closed after 5 days. New replies are no longer allowed.
