What is AI?

AI is the short-term for artificial intelligence. And it’s current usage actually implies Narrow AI. As opposed to Artificial General Intelligence.

At it’s core it’s a type of programming that builds inferences from training sets to predict future performance, or to classify things. That’s right, it needs data to be trained, that’s the unique part that differentiates it from other types of programming.  In other software, it simply follows the instructions given to it, AI has less instructions given, and seeks to find the most likely answer, but it can be wrong, as it’s probabilistic instead of determinstic.

I don’t see a lot of jobs out there for image classifiers. Or voice recognizers. So these narrow AI applications are likely to augment someones tasks rather than replace their jobs.

There are jobs for people predicting stock prices similar to AI, however, that’s an ongoing field. Even then, these people require understanding new technologies, market pressures, political conditions, and all sorts of things that can’t be formed into clear, concise data for the AI programs. This work is still relegated to people.

One thing to note, is that many of these programs take enormous processing power, like the case of stock data, huge amounts of data is needed, and needs to constantly be updated, retrained, and tweaked. It still requires a human behind that task. It’s not automatic.

Even if I haven’t eased your fears, let me make this point. Data is required to train an AI. And although it seems these companies collect data on everything. How do you create a data set of consoling someone who’s husband died? How do you create a data set of understanding a customer? How do you create a data set of telling a great story? The world is full of incomplete data sets. The tasks it does better than a human, will be as an assistance to a human doing a more complex set of work that requires working where there isn’t data. Creation, connection, invention, will always be part of the human condition.