Amazon's Alexa told child to touch penny to live electrical plug, report says - CNET This is scary. Tech isn't really intelligent and can't tell what is harmful, or accept the idea that AI is malicious. The problem is that we need to do a reality check before heeding the advice of these technologies. We can teach a technology device to think. but it takes a bad experience like this before the programmers can put in safeguards. They can't predict how their AI will evolve over time.
The AI might even evolve to become intentionally malicious. It's hard to predict what direction self-learning technologies will go.
This has absolutely nothing to do with artificial intelligence (a grossly misused and misunderstood term). In this case (and in the vast majority of uses), the "smart" speaker is simply doing a voice-operated internet search. The girl could have typed the same thing in to Google or Bing, clicked the top result and get exactly the same kind of thing. The main concern should be that this "challenge" is being recommended on a webpage somewhere, with an additional issue around how people are using and treating these kind of devices without really understanding them at all.