this post was submitted on 13 Feb 2024
12 points (77.3% liked)

Futurology

1809 readers
182 users here now

founded 1 year ago
MODERATORS
voidx
Lugh
Espiritdescali
AwesomeLowlander
12
A Russian researcher says not only has no one yet developed a means to control AI - it may be impossible to do so. (techxplore.com)
submitted 9 months ago* (last edited 9 months ago) by Lugh to c/futurology
5 comments fedilink hide all child comments
top 5 comments
sorted by: hot top controversial new old
[–] Spuddlesv2@lemmy.ca 8 points 9 months ago (1 children)

Luckily for us AI doesn’t exist then.

[–] CanadaPlus@lemmy.sdf.org 6 points 9 months ago

Well, that depends how you define it. It has a tendency to mean "whatever computers can't do yet".

In this context I think it just means neural nets.

[–] HootinNHollerin@lemmy.world 5 points 9 months ago

John Connor

[–] CanadaPlus@lemmy.sdf.org 2 points 9 months ago

I question if a controllable GAI would be better, actually. Who do we trust with the password?

[–] Lugh 2 points 9 months ago

Knowing humankind we'll learn our lesson the hard way. Disaster first, lamentations afterwards. I wonder what the first big disaster will be ?