Zeynep Tufekci says the first AI paperclip maker to hurt humanity is making us click on ads
There is a (somewhat silly) story AI researchers like to tell about the perils of AI, where an AI machine designed to do a simple task becomes super intelligent and then devotes resources to that simple task to the detriment of humanity. A clear description of this story comes from Tim Urbans (long but excellent - recommend the series) post and his story about "Robotica" which starts like this (wayyy down the page)
“A 15-person startup company called Robotica has the stated mission of “Developing innovative Artificial Intelligence tools that allow humans to live more and work less.” They have several existing products already on the market and a handful more in development. They’re most excited about a seed project named Turry. Turry is a simple AI system that uses an arm-like appendage to write a handwritten note on a small card.”
.. and ends badly for the planet due to unintended consequences. This kind of worry has always sounded silly to me, but perhaps Zeynep Tufekci's Ted talk suggests that AI's are now programmed to show anything to make us click on ads, and this is significantly hurting human society. In many ways this is more powerful than a single futuristic AI, instead we have many large companies, working hard with AIs and large databases of information about us, following our every move not he web trying to make us click on a ad. There is no moral calculus in any of this, if they think showing you Nazi imagery will make you stay on a page a bit longer to click on an ad, then the AI will do that. The combination of AI's, major investment, a profit goal and major competition to commodities our eyeballs means that we may not even need to wait for super intelligent AI to hurt humanity, just need a few better tools to add to the human condition.
The talk is well worth a watch.