Skip to main content

Levels of Artificial Intelligence

AI is a pretty broad subject, and there's lots of things that may or may not qualify.  Perhaps it'll be helpful to arrange different things that might be considered AI in a list of levels.

Without further ado, here are my Levels of AI:

Level 1 - Helping a human do computable tasks a bit faster

We've been at this level for decades (if not longer).  I think of this as just making things we do a bit easier. Like a pocket calculator - I can't personally multiply 103 by 61 in my head, but I can work it out on paper - and there are people who can do it in their heads.  A pocket calculator can be used to do this easily, so it's artificially more intelligent than me in that specific task.

This is just computation though, and I reckon a modern laptop running a spreadsheet is also Level 1.  Not really intelligent by itself, but can help us appear so.

The answer is 6,283 by the way.

Level 2 - Helping humans with their lives in meaningful ways

Today's consumer AI technology might be this level - Siri being able to organise your day by telling you what meetings you have, and the weather and so on, responding to natural language appropriately etc. Chatbots, Google Now, Siri and Alexa and similar technologies are all level 2.  I would consider IBM Watson, Google Alpha Zero and other modern machine learning systems as level 2.

Level 3 - Being able to comprehend *all* human output

What do I mean by that?  We're making attempts in this direction already - being able to recognise what's in a photo, or translate spoken language is in this area.  But what about a little boy saying "hat" and pointing at the door?  His mother might know that means he wants to go outside - and knows he has to wear his hat.  Could an AI system ever understand that? For level 3 we need a computer to be able to understand anything that a human could.

This also encompasses an AI being able to read all human text - every book ever produced; all TV and Film productions, news and YouTube.  They should be able to intelligently summarise any subject based on this knowledge.

So this level includes some comprehension - the AI should be able to comprehend some deeper meaning in the things it sees. Level 3 doesn't need to be able to create it's own material though.  That's a Human ability.  note - I totally don't count those music generators that general pleasant but boring music.  They're just a step above a random number generator.

Level 4 - Human Intelligence

Pretty scary, but from my arbitrary scale we're just 2 steps away from building Artificial Human-level intelligence. Hold on there cowboy, that doesn't mean it's imminent. Level 3 is bloody hard and even if we have a good level 3 for 20 years, we may never get to level 4.
Humans can do many many things that a Level 3 AI cannot.
  • Empathy & Sympathy.  An AI can't detect if you're feeling sad - and can't feel sadness. There are systems that can recognise some emotions based on your face, but knowing how it feels to lose a loved one or recognise embarrassment in others is something no AI can currently do.
  • Tell a good joke and know why it's funny - eg: What has four legs and one arm? A happy pit bull....  Why's that funny?  Would an AI ever get it?  They'll need to.
  • Be Creative.  Can an AI ever think up novel approaches to unique problems?  Can it use it's experience to work out a better solutions to a problem?  Very human activity right there.
How might be create a Level 4 intelligence?  Research into the Human brain might deliver abilities here.  Or perhaps running human brain simulations on powerful computer hardware. We're still pretty far away from any of these things.  Good because next we have:

Level 5 - Greater than human intelligence.

Once we move from a computer that can think like a human, it's only a matter of time until we find ourselves in a world with Super Intelligence.  There's no way we can predict what this will be like - any more than an ant or spider can comprehend our intelligence.  We probably won't have the mental capacity.



Comments

Popular posts from this blog

Chatbots might take over the world someday.

I love chatbots - I remember trying to write one when I was at university in 1996 or so. I think I used Modula-2. Anyway, It was pretty crap, but my friends and I had fun making it respond to all queries like it was drunk.  It responded to every query with either "I love you man, you know that?", "I just wish I could find someone to love." or "You wanna take this outside sunshine?". I'm pretty sure it would have passed the Turing test with all that. Although modern chatbots like Cleverbot are pretty damn good, but they haven't yet passed the Turing test as far as I know. In case it's not clear, a chatbot is basically bit of software that's programmed to respond intelligently to specific textual queries and statements.  Like if you say to an online chatbot "Hey chatboi, what's the weather in Moscow today?" it will say "Bloody cold.  You should wear a fur coat and one of those Russian hats with the ear flaps...