Jerry Kaplan Fellow and Visiting Lecturer, Dept. of Computer Science, Stanford University
to recognize with respect to Artificial THE MOST IMPORTANT THING Intelligence is that it is a technology, not a market. You can’t ‘sell AI’ (at least, not to consumers); it has to be appropriately incorporated into products and services. With the recent progress in machine learning, the most productive places to apply it are to any industry or problem for which there is a lot of data available, but not in a standardized form. AI works best on well-defined problems where success or failure are easy to measure objectively.
I’ve spent most of my career using AI technology to build real products for real customers. Over time, I could not help but notice a widespread mythology: that we are building increasingly-intelligent machines that may ultimately surpass human capabilities — and wreak havoc. In my view, this narrative is misguided and counterproductive. A more appropriate framing is that AI is simply a natural extension of our long-standing efforts to automate tasks — which date back to at least the start of the industrial revolution.
To state the obvious, machines are not people, and there is simply no persuasive evidence that they are on the path to becoming ‘generally intelligent’ — despite what we see in the movies. ‘Artificial general intelligence’, at least so far, is really little more than a pipe dream. ‘Wait a minute’, you might say: ‘doesn’t the new wave of AI technology solve all sorts of complex reasoning and perception problems?’ Sure — it can perform some tasks that people solve using human intelligence; but that doesn’t mean machines are intelligent. It merely means that many tasks that we thought required general intelligence are, in fact, subject to solution by much more narrow and mechanical means.
Robots aren’t coming to take our jobs. They’re coming — but not exactly for our jobs. Machines and computers don’t perform jobs: what they do is, they automate tasks. Except in extreme cases, you won’t roll in a robot and show an employee to the door. In- stead, what the new technology does is, it hollows out and changes the jobs that people perform. Even experts spend most of their time doing mundane repetitive tasks, like reviewing lab test results, drafting simple contracts, filling out paperwork, etc. If your job involves a narrow, well-defined set of tasks — and many do — then indeed, your employment is at risk. But if you have a broader set of responsibilities or, importantly, if your job requires a human touch — expressing sympathy or providing companionship — I really don’t think you have much to worry about.