A recent article caught my attention on speech recognition now being better than human performance. Of course, this is for a specific test, in specific circumstances, but the “better than human” marker is going to start to be broken over and over again in many ways.
It was almost 10 years ago that IBM Watson played against Jeopardy champions and just creamed them. Watson was better at parsing the questions, finding the queries, weighing the odds, and pressing the buzzer. Looking back, this win seems like a no brainer. Watson wasn’t even using speech to text in handling its responses. Other games such as Go have been bested by computer players.
What ever single dimension you use to define human intelligence can or will be surpassed by a computer. Even creativity can be challenged through GANs.
The big question is where does that leave us? More directly — are we then useless?
If a machine sews buttons better than us, it’s not the machine that benefits. If a computer composes music better than humans, it’s humans who benefit. If they write more clearly than we do, we benefit. We get to understand the world better.
Intelligence doesn’t mean autonomy. It also doesn’t mean sentience. Machines will outperform us in many ways but we are going to benefit.