Chinese Researchers Rate AI IQ

[FrgMstr]

This is a fairly long Chinese paper that delves into figuring out the IQ of different forms of Artificial Intelligence. Great fun since these guys have included the most popular versions of AI that most of us likely interact with in Google's Assistant and Apple's Siri. Interestingly they show Google's Assistant coming in at 47.58 IQ, which is right behind a six year old human (55.5 IQ). Apple' Siri checked in at 23.94 IQ.


Using this model, we observe three challenges, namely, expanding of the von Neumann architecture; testing and ranking the intelligence quotient (IQ) of naturally and artificially intelligent systems, including humans, Google, Microsoft’s Bing, Baidu, and Siri; and finally, the dividing of artificially intelligent systems into seven grades from robots to Google Brain.

 

[Ultima99]

Clearly these Chinese researchers used their white male bias to discriminate against Siri, who identifies as female. If Siri were a man she would have scored better...

 

[travisty]

That paper is pretty out of date and their conclusions about AlphaGo are incorrect. A second version of AlphaGo didn't use any human players to train from. It only played against itself. End result was a system that beat the original AI 100 to 0.

I'm curious what IQ rating Google has now in 2017.

The part about the new AlphaGo is that, in my mind, it shows that humans have been playing Go incorrectly or at least non-optimal strategies for millennia.

The other fun part is in about a years time a deep learning system went from the best grand master Go player to basically an impossible to beat Go prodigy. Exponential growth in action!

If computers are already at a new-born's IQ level, the distance to go to super-human is less than the time it took to go from an insect's iq to its current iq.