And now for something completely meta. We’ve just about got used to the idea of AI understanding, learning – possibly even creating – but actually forming opinions? To go one further, here a robot debates the case for robots. Against itself. As this article from NewScientist asserts, the AI still relied on human inputs. However, the key development is that it was able to understand the arguments, perhaps signalling awareness of its own existence alongside humans?
A robot has visited the University of Cambridge Union to debate the dangers of artificial intelligence. Against itself. Project Debater, developed by IBM, combined arguments from more than 1,000 humans to discuss the proposition: ‘AI will bring more harm than good’.
Using these online submissions, Project Debater analysed them to identify themes and eliminate redundancy based on repetition – even catching when opinions expressed the same point using different words.
The AI was able to present the two arguments, albeit with a few repeats and an occasional lack of examples.
Despite that, the opposing side, in support of AI’s benefits, won 51.22% of the audience vote. Project Debater’s first attempt earlier this year against champion debater Harish Natarajan was not successful.