On one hand artificial intelligence has the potential make our lives easier, although it is susceptible to picking up some of the less desirable aspects of humans as studies of bias have shown. At the other end of the spectrum there are the risks, beloved of sci-fi authors, of super-intelligent machines that calculate that there is no need for humanity. In-between this there is a range of different outcomes that will be reliant upon those developing artificial intelligence to be mindful of the risks.
In discussing these challenges, science news site Risk Bites has made a video outlining the ten potential risks of artificial intelligence that everyone in society should be thinking about as the technology progresses.
The ten risks listed address (with the video playback timings in parentheses):
1. Technological dependency; (1:07)
2. Job replacement and redistribution; (1:25)
3. Algorithmic bias; (1:43)
4. Non-transparent decision making; (2:03)
5. Value-misalignment; (2:27)
6. Lethal Autonomous Weapons; (2:44)
7. Re-writable goals; (2:59)
8. Unintended consequences of goals and decisions; (3:11)
9. Existential risk from superintelligence; and (3:31)
10. Heuristic manipulation. (3:51)
These risks represent some of he most important questions that need to be considered together with the nature, context, type of impact, and magnitude of impact of the risks. These risks can be weighed up together with relevant benefits and trade-offs.