Is Google putting translators out of business?

Google recently updated its popular Google Translate tool, and in doing so, claimed the new version was almost as accurate as human translators. Before this development, Google Translate and other automated tools for converting one language into another were long more of a source of mirth than of accurate translations. Language is one of those things that computers are thought to be particularly bad at. It relies on fluid rules, intuition, taste, and context, so much more than, say, raw number-crunching.

But this claim that a robot can do something almost equally as well as a human raises a seemingly eternal question: Will robots replace us? Will they take all our jobs, leaving us destitute in a post-post-industrial dystopia where we all cash in our basic income checks while playing Xbox and eating (drone-delivered) pizza all day?

Take comfort in this: So far, there's no such thing as "general" AI, meaning artificial intelligence that can think and do all the tasks a human does. And there probably won't ever be. What we have instead is AI that's increasingly sophisticated at very narrow tasks. They augment and assist in human work, but they don't replace it.

Consider the security software and consulting company Palantir. It was founded out of precisely this insight: that humans and computers are complementary. Palantir grew out of the anti-fraud work at PayPal, when fraud threatened to consume the company. It was impossible for humans to review every transaction, but automated fraud detection systems were much too dumb to be effective. PayPal managed to keep fraud down by combining the work of computers and humans. Computers flagged suspicious transactions and humans reviewed them. Palantir then applied this logic to counter-terrorism in the wake of 9/11, and online security more generally.

Palantir's software and consulting doesn't try to use AI as some sort of magic bullet to fix every problem, but instead uses AI to make it easier for human analysts to find suspicious activity.

This is just one example of a broader phenomenon. AI apps are still narrowly focused and probably serve more to augment our jobs, rather than to replace them.

A recent McKinsey & Company report on workplace automation came to a similar conclusion. The report has a clever angle: Instead of looking at "occupations" (like "electrician" or "psychiatrist"), it breaks down each job into a number of "activities" (for a retail salesperson, "demonstrate product features" and "process sales and transactions") and looks at which of those activities are likely to be automated.

The report finds that while many activities will be automated, few occupations will be. In other words, robots won't make most of the jobs disappear, but they will make those jobs very different, and in many cases, easier. What's more, the impact will be felt across the board:

"We estimate that activities consuming more than 20 percent of a CEO's working time could be automated using current technologies. These include analyzing reports and data to inform operational decisions, preparing staff assignments, and reviewing status reports. Conversely, there are many lower-wage occupations such as home health aides, landscapers, and maintenance workers, where only a very small percentage of activities could be automated with technology available today..." [McKinsey]

What activities will benefit most from this shift to automation? Things like creativity, and sensing emotion, according to the report. Think about health, for example. Already, computers are often better at diagnosis than human doctors. But tomorrow's health professionals will be able to spend more time getting to know their patients and coaching them through their options than performing diagnostic work that can easily be done by AI. And this is better for everyone.

So, fear not. We won't lose the Earth to our robot overlords. At least not anytime soon.