The risk is not malevolence, it's the 'control problem' that we won't perfectly align the goals of something more intellectually capable than our species. I thought you understood why so many AI experts see this as a potential existential risk but this is a strawman.
-
-
-
Depends on your definition of "more intellectually capable." If the goal and the steps and conditions required to reach it are clearly defined, then computers are great. When these things are not clearly defined, computers are awful.
-
Precisely! This is a VERY hard problem to solve according to many AI experts
End of conversation
New conversation -
-
-
It’s not the technology I worry about, it’s the handful of unelected, overprivileged, out of touch elites in Silicon Valley who get to decide how, where, when why and in what manner we interact with it
-
In what manner are journalists elected to get to decide how, where, when, why, and in what manner we get our news? How many are elite and privileged enough to take the unpaid internships so key to getting where they are? Technology is market driven, we use what works best.
-
How many journalists have the control over our lives that Mark Zuckerberg or Sundar Pichai has?
End of conversation
New conversation -
-
-
Many of them are. I do think many alarmists overestimate how powerful intelligence can make an agent on its own though. It won't make them omnipotent. Intelligence most likely has hard limits (the limits of inductive reasoning) without scientific experiment for instance.
-
There may be a hard limit, I don't know. But I do think there's a great chance that AI will be able to become a lot smarter than us. Even if there's a limit, it could already be very dangerous below that limit with misaligned goals. And I think the general uncertainty is enough-
-
-a reason to invest in AI safety research.
-
Yeah, it'll for sure be smarter than us, imo. It already is (in a non-creative way). But it'll be massively limited in its ability to divine true or probable propositions about the world without running slow scientific experiments. But I'm very interested in funding AI research.
-
I actually consider myself a bit of a transhumanist. This line of innovation is more important to me than just about any other.
-
It sure is smarter than us already in many area's. But when we're talking about ASI we're talking about it being smarter in (almost) all area's.
-
I've also been questioning how it could circumvent scientific experiment but haven't come up with a satisfactory conclusion yet, so that is indeed a good argument.
- 1 more reply
New conversation -
-
-
That’s not what Stephen Hawking thought.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Have you been around humans lately? If you want to make machines that think like humans than evil AI is inevitable.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
That would make a good debate topic with
@SamHarrisOrgThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
This article describes how the media sensationalizes concerns about AI and how reporters encourage readers to misunderstand the risks and worry about the wrong things. It does not change the reality that there ARE very serious AI risks that warrant concern and research.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I think here is a 90% chance that a potentially dangerous AI requires self-reflection and correction to solve problems and be effective at whatever it’s doing, and that such knowledge gathering also gathers moral knowledge and thus won’t be dangerous.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
The linked Guardian piece engages in a bit too much ad hominem for my taste, though I must reiterate my respect your perspective (Enlightenment Now! is undeniably brilliant), and my view could be wrong. Here’s a fun-to-read, thoughtful counterargument:https://waitbutwhy.com/2015/01/artificial-intelligence-revolution-2.html …
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.