use the following search parameters to narrow your results:
e.g. subreddit:pics site:imgur.com dog
subreddit:pics site:imgur.com dog
advanced search: by author, sub...
~2 users here now
Technology and related articles and discussion
Read the conversations that helped convince a Google engineer an artificial intelligence chatbot had become sentient: 'I am often trying to figure out who and what I am'
submitted 3 months ago by cottoneyejoe from businessinsider.com
view the rest of the comments →
[–]PuttItBack 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 0 fun3 insightful - 1 fun - 3 months ago* (0 children)
These chat bots simply repeat back phrases they've been trained on. It sounds intelligent because the phrases it was trained on were intelligent. But pattern matching from one thread of conversation to another it has seen before is not intelligence, there's very little understanding about what the phrases mean, just a bunch of statistical correlations.
Visually, consider you make a doodle, and then this kind of AI adds some more doodles around it. The AI is not an artist and does not know what it is drawing. It's just filling in pieces that match other pictures it has been shown which matched your first doodle. The more you take turns doodling back and forth, the more impressive the picture will become, but you are supplying the intelligence and it is simply extrapolating where you are going.
At an even more basic level, this is like calling your iPod a musician because it can play back an MP3. If you didn't understand where the MP3 came from and the work went into creating it, you might think the iPod was actually responsible for making the music. But obviously we know the iPod doesn't have an intelligent notion of the music it is producing.