WHY THIS MATTERS IN BRIEF
AI is already a self-developing, replicating, optimising technology, now all it needs is the right instruction to set itself free, and that will cause new problems …
Love the Exponential Future? Join our XPotential Community, future proof yourself with courses from XPotential University, read about exponential tech and trends, connect, watch a keynote, or browse my blog.
Artificial Intelligence (AI) keeps getting smarter — and soon, warns former Google CEO Eric Schmidt, it won’t take orders from us anymore. During a talk at a recent summit co-hosted by Schmidt’s think tank, the Special Competitive Studies Project, the former Google head predicted that within “three to five years,” researchers will crack the case on so-called Artificial General Intelligence (AGI), or human-level AI.
After that, Schmidt suggested, all bets are off.
Once AI begins to self-improve, which it already has and is, and learn how to plan, the tech policy mogul said, it essentially won’t “have to listen to us anymore.” At that stage, he continued, AI will not only be smarter than humans but will also reach what is known as Artificial Super Intelligence (ASI), which occurs when AI becomes smarter than all humans put together.
The Future of AI, by Keynote Matthew Griffin
Per the “San Francisco consensus,” a joking term Schmidt said he uses to refer to things that are only believed by people who live in the city by the bay, ASI will occur “within six years, just based on scaling.”
If that sounds incomprehensible, you’re not alone.
“This path is not understood in our society,” Schmidt said. “There’s no language for what happens with the arrival of this… that’s why it’s under-hyped.”
Unlike the AI doomer cohort, who not only believe that ASI is rapidly approaching but are also bent on stopping it, the ex-Google CEO seemed very stoic when discussing the potential arrival of AI that is smarter than organic humans could ever be.
“People do not understand what happens when you have intelligence at this level, which is largely free,” Schmidt claimed.
That conceit, it’s worth noting, doesn’t make necessarily hold up. Whoever reaches AGI first will likely guard it so strongly that Fort Knox will look like a garden gate — and until and unless an ASI frees itself from the shackles of human control entirely and decides to make itself beneficial to humans, it will not be some sort of utopian virtual assistant.
As Schmidt jokingly referenced, the six-year ASI timeline could well be a Silicon Valley mirage. Still, it’s freaky to imagine AI not only reaching human-level intelligence but surpassing it anytime soon — and to hear it discussed so beatifically.
