http://www.faz.net/-gqe-94jtq

FAZ.NET-Talk with Max Tegmark : „Humanity might flourish like never before“

Max Tegmark Bild: dpa

Physicist Max Tegmark thinks a lot about the impact of technology on society. He explains chances and dangers of AI, biological intelligence - and why he’s an optimist about progress in that area.

          Max, is AI dangerous?

          Alexander     Armbruster

          Redakteur in der Wirtschaft.

          Even fire is dangerous, but that doesn’t mean we shouldn’t use fire to warm our homes. The more powerful the technology, the greater its dangers and its benefits. I’m optimistic that we can create an inspiring future with AI if we win the race between the growing power of AI and the growing wisdom with which me manage it.

          Stephen Hawking recently reinforced his big warning: „I fear that AI may replace humans altogether“, he said. That’s a gigantic exaggeration with respect to where we are at the moment as even experts don't know a clear path to anything like an AGI.

          We don’t have AGI at the moment and don’t know how to build it, but recent surveys show that most AI researchers think we’ll get AGI within decades. So it’s perfectly reasonable to talk about its risk. For example, the Asilomar AI Principles mention superintelligence and existential risk and are signed by over 1,000 AI researchers, including industry leaders from Google, Deep Mind, Apple and more.

          Most of what is behind Machine Learning, Deep Learning or Artificial Neural Networks are classification-issues and very sophisticated statistics – nothing to worry about.

          Even today’s low-level AI poses interesting challenges involving job automation, cybersecurity, voter manipulation, lethal autonomous weapons, etc. But the greatest challenges – and opportunities – arise if the AI field succeeds in its original goal of outperforming all aspects of human intelligence. 

          One of the problems of the discussion about AI is that there are different understandings of what intelligence really is. Marvin Minsky ones called it a portmanteau – everybody kind of puts in what she wants. Do you have a good definition?

          I define intelligence simply as ability to accomplish complex goals. I give this broad definition because I despise “carbon chauvinism”, the arrogant attitude that things can’t be intelligent if they aren’t made of carbon.

          Isn’t intelligence a concept, that in our common understanding is strongly aligned with a biological organism?

          We've traditionally thought of intelligence as something mysterious that can only exist in biological organisms, especially humans. But from my perspective as a physicist, intelligence is simply a certain kind of information processing performed by elementary particles moving around, and there's no law of physics that says one can't build machines more intelligent than us in all ways. This suggests that we've only seen the tip of the intelligence iceberg, and that there's an amazing potential to unlock the full intelligence that's latent in nature and use it to help humanity flourish.

          Even if there is a kind of superintelligence one day, why should that be an essential problem for us? I’d like to do a comparison: The Industrial Revolution did neutralize physical strength, but today nobody cares, that cars can drive faster than the fastest human can sprint.

          What makes us humans the dominant entities on this planet isn’t our strength but our intelligence, so if we build machines much smarter than us, there’s no guarantee that we’ll stay in control. Moreover, malicious humans could use AI to dominate other humans. Then there’s the near-term issue of jobs: when the machines of the industrial revolution outcompeted our muscles, we educated ourselves to get better jobs using our brains. If machines of the AI revolution outcompete our brains as well, there’ll be no new jobs we can do more cheaply than machines.

          In your new book, you introduce the term „Life 3.0“. What do you mean by that and in contrast to that, what is „Life 2.0“ and „Life 1.0“?

          I call bacteria „Life 1.0“ because they’re really dumb, unable to learn anything during their lifetime. I call us humans „2.0“ because we can learn, which we in geeky computer jargon can view as installing new software into our brains – like when I chose to study German in school in Sweden. „Life 3.0“, which can design not only its software but also its hardware, still doesn’t exist – but we seem to be heading in that direction. Perhaps we should call ourselves „2.1“ now, since we can install artificial knees, pacemakers aned cochlear implants.

          What should we as a society do to prepare for the advance of smarter computers?

          First: Prevent an arms race in lethal autonomous weapons, which the UN is discussing. Second: Ensure that the huge wealth at AI will produce is shared to make everyone better off rather than impoverishing and alienating large fractions of the population. Third: Invest heavily in AI safety research. How do we transform today’s buggy and hackable computers into reliable AI systems that we can really trust? How do we make machines understand, adopt and retain our goals?

          What can every individual do? Learn coding ourselves or at least enable our children to learn it very early?

          Go into professions that machines are currently bad at, and therefore seem unlikely to get automated in the near future, for example jobs involving unpredictability, creativity and social intelligence. Software development makes up only 1 percent of the job market, so it can’t be the solution for most people.

          Are you an optimist?

          I’m optimistic that we can help humanity flourish like never before with advanced AI. But this won’t happen automatically, like the Sun rising over Germany tomorrow – we’ll need to plan and work hard to ensure that we use this powerful technology wisely. One final comment: Everything I love about civilization is the product of intelligence, so if we can amplify our human intelligence with AI and solve todays and tomorrows greatest problems, humanity might flourish like never before.

          Quelle: FAZ.NET

          Weitere Themen

          Hier können Sie die Rechte an diesem Artikel erwerben.

          Topmeldungen

          Senats-Wahl in Alabama : Amerika ist so gespalten wie nie zuvor

          In Alabama haben fast drei Viertel der weißen Männer und rund zwei Drittel der weißen Frauen für den umstrittenen Roy Moore gestimmt – die Afroamerikaner gaben zu 90 Prozent dem Demokraten Jones ihre Stimme. Nie war der Graben tiefer, der mitten durch Amerika geht.
           Nur nicht durch die Decke gehen: Die Wintersport-Experten des ZDF für die Olympischen Winterspiele in Pyeongchang.

          ARD und ZDF haben Geld : Steigt der Rundfunkbeitrag doch nicht?

          Die Finanzkommission Kef hat die Finanzen des öffentlich-rechtlichen Rundfunks für die Jahre 2017 bis 2020 unter die Lupe genommen. Sie kommt zu dem Schluss: Da lässt sich einiges sparen.

          Newsletter

          Immer auf dem Laufenden Sie haben Post! Abonnieren Sie unsere FAZ.NET-Newsletter und wir liefern die wichtigsten Nachrichten direkt in Ihre Mailbox. Es ist ein Fehler aufgetreten. Bitte versuchen Sie es erneut.
          Vielen Dank für Ihr Interesse an den F.A.Z.-Newslettern. Sie erhalten in wenigen Minuten eine E-Mail, um Ihre Newsletterbestellung zu bestätigen.