Computers talking to each other

There is a lot of development going on in artificial intelligence. AI is back from the long “AI Winter”, and this time it seems like it is getting somewhere useful. There are types of programs which we would’ve previously called AI, the we now sometimes called the machine learning, doing real valuable work.

(How soon will this end up with a real genuinely “intelligent” computer program? Is the singularity near? Is it far? I have no idea.)

But I think this is how actual AI will make it into our lives. In an irritating, mundane, wasteful way.

Our smart phones are getting smarter and smarter. There are now features by which they can help us answer incoming messages, SMS and email. They can gather statistics about the way we have answered different messages, they can use all of the vast trove of data that are gathered about all of our activity (hello Google, how are you today?), they can mix in some “AI” techniques, and produce much-better-than-random results.

Now simply imagine progress on this for a few years, with faster and better processors in our smart phones and in the cloud. Eventually, our smartphones will get pretty decent at offering us proposed replies to incoming messages, guessing what we would say to each other. Eventually we will get tired of having to always tell our phone “yes, go ahead and answer the way you guessed”. We will be a switch offered which simply turns that feature always-on. Our devices will then be able to answer incoming messages, automatically, as if they are us, without our intervention.

More than one person will turn this on.

Artificial intelligence will then consist of our devices talking to each other, pretending to be us. If this works well enough, are we even needed at all? Some days it seems like all I do is read and write email. On such a day, if my devices could guess what I would’ve written sufficiently well (not perfectly, just sufficiently well), I can simply take that day off. Or that week. Or forever.