Log in

No account? Create an account
Musings On Artificial Intelligence and Seeming Intelligence [Notes on the Acceleration] - Synchronicity swirls and other foolishness

> Recent Entries
> Archive
> Friends
> Profile
> my rpg writing site

April 11th, 2010

Previous Entry Share Next Entry
03:14 am - Musings On Artificial Intelligence and Seeming Intelligence [Notes on the Acceleration]
True (meaning conscious) artificial intelligence remains a vast unknown. Perhaps we'll invent it in a couple of decades, perhaps it will take longer, or perhaps it's impossible. We have no idea. The newly invented memristors may prove a path to AI, or they might be another in a long line of dead ends. Perhaps brute force emulation will do the trick - we'll know in perhaps 15-20 years, when Moore's law gives up the power to do real time brain simulations. Here's a few more thoughts to consider. Work is being done, but no one has any clue what or when the results will be.

However, something else is definitely coming quite soon - programs that can speak and write like humans, and presumably eventually interact like them in real-time. We're just seeing the edges of this, and I don't believe that these programs need be actually intelligent, but humans will be able to listen, talk, and and interact with them as if they were people, and that's very interesting. If they seem sufficiently like people and are usefully interactive, then how we deal with software, and with the many robots we will soon have all around us in daily life will be very different from how people deal with most technology now.

Even if true artificial intelligence proves impossible, we're going to have people interacting with seemingly programs and robots in less than 15 years. It will take a while for the software to be totally convincing, but for most causal & brief interactions with clerks and service personal (or in this case service robots), I can see the first fairly impressive, if also obviously imperfect results starting in 5-7 years. Already, typing search queries into google as complete sentences is considerably more useful than it was 5 years ago - I'm betting that by 2020, many people will be having useful text and voice conversations with software and robots. We shall definitely see.
Current Mood: thoughtfulthoughtful

(1 comment | Leave a comment)


[User Picture]
Date:April 11th, 2010 02:32 pm (UTC)
This could have short-term negative consequences, as many customer service and retail workers would be thrown out of their jobs.

However, I could also see the enormous potential of robots for social skills training. People with Asperger's syndrome, among others, could interact with robots designed to respond like actual people. By practicing appropriate interactions, people with autistic spectrum disorders can improve their ability to secure employment, friends, and relationships. (BTW, I have been diagnosed as being on the spectrum myself, so I am not saying this out of a desire to marginalize or mock others. I wish I had a robot to practice conflict resolution skills with, so I can know when it is appropriate to defer to someone else and when it is not.)

> Go to Top