AI Risk #2 - Overtaking Humans

With the spurt in what AI can do, I re-read Nick Bostrom’s Superintelligence. In the intro, he had written:

“The control problem – the problem of how to control what the superintelligence (aka AI) would do – is quite difficult.”

Even more ominously:

“It also looks like we will only get one chance. Once unfriendly superintelligence exists, it will prevent us from replacing it or changing its preferences.”

 

Some feel this is an excessively pessimistic view. Won’t an AI rise be gradual, allowing us time to formulate and tweak our response? Not necessarily, argues Bostrom. Why not?

“The (AI) train might not pause or even decelerate at Humanville Station. It is likely to swoosh right by.”

In other words, AI might just explode, growing exponentially. It would hit human level abruptly and then continue on its upward trajectory at that same exponential pace, leaving us no time to react.

 

Now keep in mind Bostrom wrote his book back (it feels a lifetime ago) in 2013. Yet, even then, he was able to correctly anticipate many things.

 

There is a field of probability called Bayesian theory – simply put, it tells us how much we should adjust our belief system when we get some new info. By how much should the new data strengthen our belief? Or conversely, by how much should our confidence in our belief decrease? This is at the heart of how AI’s “learn” – they consume more and more data and use it to adjust their “understanding” of things. Bostrom correctly anticipated that improvements to Bayesian algorithms would “yield immediate improvements across many different areas”. I feel this is why we find AI’s got so good at language (ChatGPT) as well as art almost simultaneously.

 

Worryingly, he wrote:

“If somebody were to succeed in creating an AI that could understand natural language as well as a human adult, they would in all likelihood also either have succeeded in creating an AI that could do everything else that human intelligence can do, or they would be a very short step from such a general capability.”

With ChatGPT, the barrier called language has been breached.

 

When would an AI be deemed to exceed human intelligence? Bostrom’s take:

  • It would exceed cognitive performance of human in virtually all domains of interest;
  • It would be able to “learn” as it goes along;
  • It would know to deal with uncertainty and probabilistic knowledge;
  • It would be able to extract useful concepts using data from its sensors + its internal states;
  • It would know how to improve its own architecture.

Today, AI has checked all of the above, except the last point – its architecture is decided by us humans, both in hardware and software. But with AI starting to write software already, how long will it be before it can write software better than us and start to design hardware ideal for its progress?

 

Which brings us back to the urgency of Bostrom’s “control problem” from the top of this blog – and the ominous point he made that we might have only one shot at creating a control system for the AI…

Comments

Post a Comment

Popular posts from this blog

Student of the Year

Image
As a toddler, my daughter loved Alia Bhatt’s songs, esp. the ones from Student of the Year . (That movie did have many good songs and well-choreographed dances). Fast forward to present day, when she is 12 yo. She rarely watches Hindi movies now, but one of the few ones she likes of late, Raja aur Rani ki Prem Kahani has, yes, Alia Bhatt.   If I had any hopes that there was anything to be read in the name of her toddler-time favorite ( Student of the Year ), they have been dashed. Our conversations come exam time are perfectly captured by this photo and its caption: In fact, the pic below captures how she and I look at her answer sheets – me intently, she with an expression of “It’s over, so why are you even looking at it?”   Of late, we pull her leg and tell her that she isn’t showing any signs of academic prowess, she’d better find a rich guy to marry. Some time later, she showed us this meme on the Internet: “Me No Study, Me No Care Me Go Marry Millionaire If He Die
Read more

The Retort of the "Luxury Person"

Image
Coming up with biting and witty retorts. It’s something we dream of, but rarely manage to pull off. Worse, the more we stew over the offending remark, the more likely we are to come up with that dream retort, as Calvin ruefully points out: During our last holiday at the Taj Resort, seeing my wife carrying towels to our poolside chairs, one of the kids there mistook her for a member of the staff and asked for a towel. When my wife narrated this to us, I did not imagine that my 8 yo daughter would come up with her own retort. But a day later, my daughter recapped the incident, an incident that didn’t even involve her in the first place , and said, “You know what I’d have told that kid who asked for the towel?”. And without waiting for an answer, she told us anyway: “I don’t work here. I too am a luxury person who is living at this resort. Go get your own towel.” We were highly amused by the phrase she had coined: “luxury person”. It certainly captures the kind of holid
Read more

Animal Senses #7: Touch and Remote Touch

Instinctively, we think touch is based on contact. Not entirely true, explains Ed Yong in Immense World – there’s a more important aspect than contact: “These incredible feats are possible through movement. If you rest a fingertip upon a surface, you can get only a limited idea of its features. But as soon you’re allowed to move, everything changes. Hardness becomes apparent with a press. Textures resolve at a stroke.” As you run your fingers over a surface, they hit the tiny peaks and troughs and set off vibrations in the mechanoreceptors at their tips.   At the tip of its snout, the star-nosed mole has many pairs of hairless appendages. It uses them to touch things around itself. Alive or dead? Food or ignore? It does these movements at an insane speed. In the time it takes to blink our eyes, the star-nosed mole can probe, identify and swallow its food. “Light may be the fastest thing in the universe, but light sensors have their limits, and the star-nosed mole’s sense of
Read more