Welcome to Weekly AI News & Insights from Lexalytics, a curated selection of articles and interest pieces brought to you by the leaders in “words-first” artificial intelligence. This week’s AI news: AI predictions for 2018; where AI is already rivaling humans; and the Woz is an AI realist.
Reality checks and AI predictions for 2018
Not everyone is blindly optimistic. Last year, Jeff Catlin laid out some realistic predictions for AI in 2017. In his latest Forbes article, Jeff reviews his predictions and makes new ones for the year ahead.
For example, in 2017 Jeff predicted the rise of Big Brother under the guise of Big Tech – a no-brainer, there. Jeff also predicted (correctly) that AI would be forced to “consider its privilege”. Indeed, even after the lessons of Microsoft’s disastrous experience with Tay the Racist Chatbot, this past year was full of stories of biased and racist AIs.
As Jeff points out, “Any AI training data reflects the biases of our society — and the people doing the testing — and we need to improve as a society if we want AI to follow our lead.”
Where AI is already rivaling humans
Reality checks and predictions are all well and good, but how about here and now? In this KDnuggets article, Alok Aggarwal explores six areas where AI is already rivaling or exceeding human abilities.
A commonly-cited example is how Google’s DeepMind made big AI news headlines when their AlphaGo program defeated the reigning world champion of Go, Lee Sodol. But AI isn’t just about playing board games. Alok identifies more areas where AI is currently excelling: self-driving and autonomous vehicles; psychiatric diagnostics; chat-bots and personal assistants; and AI-based robotic process automation.
Steve Wozniak is an AI realist
Elon Musk is an AI alarmist. Jeff Catlin is an AI realist. So, apparently, is Steve Wozniak, co-founder of Apple. In fact, Steve says, “I agree with the A and not the I of artificial intelligence.” Ouch!
That’s from Steve’s recent interview with The Economic Times. And there’s more. “Google can look at 80,000 pictures of a dog and it may get it right sometime,” Steve points out. “Show a dog picture to a one-year-old child only once and she will recognize and know a dog forever.”
It may sound harsh, but Steve raises an important point. The simple fact is, artificial intelligence is nowhere near to approximating human intelligence – let alone exceeding it.
Weekly AI News & Insights
Enjoyed these articles, and want to discuss? Leave a comment below! Curious how Lexalytics can transform your unstructured text data into insights and money? Drop by our website to learn more.
Leave a Reply