What is the biggest AI problem? Machines are not human.
Common learning is what most people consider to be common sense, but it is often biased.
As mentioned in the video, AI’s biggest short-term problem is an excessive emphasis on data set sizes, regardless of accuracy, representation, or accountability.
The AI problem is that it will not replace humans, but rather complement them. Merging is neither necessary nor recommended.
When we stop to think about it: buying a racehorse and demanding that it behave like a camel is the same as insisting on it functioning like one. It is doomed for failure. There are only two possible outcomes: either humans are replaced, or they aren’t. If we fail, we will have failed. AI development will fail if we are not replaced.
It’s time to change direction.
It’s easy to understand how we fit into the picture when I imagine a super-intelligent learning machine that is a data junkie.
We all know that correlation does not equal causation. Except robots.