Trying to wrap your head around the the potential dangers and benefits of artificial intelligence (AI) isn’t easy. At one end of the spectrum we have AI bots that could potentially make our lives easier, and yet seem susceptible to picking up some of our attractive traits. While at the other end there’s the specter of super-intelligent machines that decide the one thing they really can’t stand is people. And in between, there’s a whole landscape of AI applications that could make our lives better or worse in in subtle, and sometimes not so subtle, ways.
To unpack some of the challenges here, Risk Bites has a 5 minute introduction to ten potential risks of AI we should probably be thinking about now:
The video explores the risks using plain language and quirky examples (spot the shout-out to 2001 A Space Odyssey, and some rather tongue in cheek references to Amazon’s Alexa). But in more technical terms, the ten risks the video addresses are:
- Technological dependency; (1:07)
- Job replacement and redistribution; (1:25)
- Algorithmic bias; (1:43)
- Non-transparent decision making; (2:03)
- Value-misalignment; (2:27)
- Lethal Autonomous Weapons; (2:44)
- Re-writable goals; (2:59)
- Unintended consequences of goals and decisions; (3:11)
- Existential risk from superintelligence; (3:31) and
- Heuristic manipulation. (3:51)
If you enjoyed the video, please support Risk Bites by subscribing to the channel, and spreading the word!