Shanahan: Superintelligent AI will naturally seek self-preservation and resource acquisition
Murray Shanahan · a16z Podcast | Artificial Intelligence and the 'Space of Possible Minds' · Jan 2, 2019 · at 39:22
Murray Shanahan, Professor of Cognitive Robotics at Imperial College London, explains philosopher Nick Bostrom's theory of convergent instrumental goals in AI.
0:00 / 0:30exact quote · 30.8s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“Anything that's really, really smart is going to have a number of goals that, that anything is going to share, and these are going to be things like self-preservation and gathering resources, if it's sufficiently powerful, then any goal that you can think of, if it's really, really good at solving that goal, then it's going to want to look, preserve itself, first of all, because how can it, you know, maximize the number of paperclips in the world, to use Nick Bostrom's argument, if it doesn't preserve itself, or if it doesn't gather as many resources as it can.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from Murray Shanahan
Shanahan: AI can learn world dynamics vicariously through web videos
“What I mean by vicarious embodiment is is that it uses the embodiment of others and to gather data. For example, the enormous repository of videos there are on the internet. There are zillions of videos of people picking up objects and putting things down and …”
Shanahan: Big corporations will dominate the future of AI
“Well, if you were to ask me to place a bet at the moment, I would place it on on the big corporation side.”
Shanahan: Reinforcement learning progress does not require massive datasets
“Actually, DeepMind are another example of the same thing, because if you want to apply reinforcement learning to games, and that's enabled them to make some quite fundamental sort of progress, you don't need vast amounts of data either.”
Shanahan: Machine learning must be embedded within larger cognitive architectures
“I see machine learning as a kind of subfield of artificial intelligence, and it's a subfield that's had tremendously a tremendous amount of success in recent years, and is going to go very, very far, but ultimately, the machine learning components have to be e…”
Shanahan: Future AI will be an ambient internet presence, not humanoid robots
“In the future, rather than the AI necessarily being the stereotype of a robot standing in front of us, it's going to be something that sort of is, is, is, Hidden away on the internet, and there's a kind of ambient presence that goes with us wherever we go.”
Shanahan: DeepMind's published DQN algorithm lacks inner rehearsal capabilities
“For the bit of work that they actually published, I think one of its shortcomings, actually, is that, in fact, although it has done all that learning about what the right action to do in, in the right circumstance is, it doesn't actually do in a rehearsal. It …”