Astro Teller on Why Artificial Intelligence Is Not Scary

“Taking over the world is an intensely human thing to want to do,” says Astro Teller, in a short interview conducted at the 2017 Aspen Ideas Festival. At Google X, Teller studies and develops artificial intelligence. Here, he argues that current frenzy over the topic might be overblown.

22 comments

  1. Nic Fishman says:

    The fundamental issue with this video is that it fails to make s distinction between machine learning and other algorithmic techniques for complex problem solving and actual Artificial General Intelligence. No one with any understanding of the field is afraid the Netflix up next recommendation system is going to take over the world; the fear is that someone will develop an AGI that might destroy the world.

  2. Norman Oro says:

    I’m not an expert on artificial intelligence, but I know a little about it; and on the whole, I share Astro Teller’s optimism regarding the technology. However, I think vigilance is definitely called for because as AI is interwoven into more and more facets of our lives and as communication between AI-enabled devices expands in order to further optimize different facets of our lives, the danger grows that the complexity of the interaction between all of these algorithms might lead us to unforeseen bad outcomes.

  3. BuonoBruttoCattivo77 says:

    The A.I may share the values of the humans that created it (think military, state security, etc). The A.I may be programmed to achieve a set goal and come to see humans (or certain humans) as an impediment (what % of the population could be deemed inefficient from certain perspectives, etc). A.I will almost surely continue to make our lives much better, until it doesn’t. A.I’s progress probably won’t be linear (could match us, then shoot past in a matter of days, hours, minutes).
    It’s not alarmist to want government and corporations to get ahead of this problem.

  4. R.D. Dragon says:

    Not an argument. “Obviously a robot would never do anything unexpected so you’re crazy to be worried.” The more things we put AI in control of (including weapons), and the more complicated we make its decision processes, the more opportunity there is for the AI to surprise us and the greater the consequences. When very smart and informed people like Elon Musk are worried about this issue, I’m inclined to take it seriously.

  5. Peter Belanger says:

    Here’s why artificial intelligence is scary, the effect it is going to have on us humans.
    Let’s call it ‘the calculator effect’. People have already begun losing basic arithmetic skills as pocket calculators have become more common. Let’s not forget the spell checkers as well, many of us honestly can not spell anymore, we have become dependent on spell check to help us. Turn it off and watch yourself struggle.
    As machines start doing their tasks for themselves, then we humans will just go along with whatever the A.I. suggests.
    Humans are lazy, we all now that. If we have the machines to do all the thinking for us, then what do we have do think about?
    At best, I see a “Wall-E” future in this. Better start working on your super-comfy, all-in-one entertainment chair with the 144oz drink holder now.
    ….Now that’s scary!

  6. bostero2 says:

    The video mixes up automation with AI. They are not the same thing, machine automation follows a program designed by a human to perform a specific task, if anything is wrong with the programming it needs to be reprogrammed. AI has self-learning, if anything is wrong with the programming it will try to fix itself. There is no AI in your car, you have automation technology which allows you to delegate certain tasks (and presumably the whole of driving) but it doesn’t learn on its own, it doesn’t think, it computes and follows a program that specifies what it should do on any particular occasion.

  7. Nicktoon Maniac says:

    Look, if you start Autonomous weapons, they’ll get smarter and try to kill us, so the best way for robot weapons is if you have them radio controlled. Also, can we have driverless cars be cancelled so crash proof cars can replace them so people who cannot drive make their lives easier?

Comments are closed.