The night of Thursday 21 November saw a momentous moment happen for the 200-year-old Cambridge Union debating club. For the first time ever, an AI robot participated in the club's debate.
This robot was IBM'sProject Debater AI software. It was used to assist two teams of human debaters as they discussed the issue of whether or not artificial intelligence will do more harm than good.
There were some deeply ironic moments when the AI machine was sharing information about how much harm AI could impose.
RELATED: ELON MUSK SHARES HIS VIEWS ON AI, NEURALINK, AUTOPILOT, AND THE BLUE DOT IN A PODCAST
What is the Project Debater?
IBM created the Project Debater as a software system than can collect and categorize arguments from text or audio information, then summarizes the positions into a smooth synthesized speech.
The reason behind Project Debater, and using it as part of the Cambridge Union debate, was to showcase its capabilities ahead of IBM offering it to the public. Called 'speech by crowd', Project Debater collects and summarizes huge amounts of arguments sent by individuals.
For instance, before this particular debate, the machine went through over 1,100 disparate arguments about whether or not AI would do more harm than good that people sent through to IBM's website.
570 arguments were categorized by the machine as pro-AI, and 511 arguments were opposed to it. Some other comments were discarded as they were deemed irrelevant.
Using a synthesized woman's voice, Project Debater was used to present the opening case for each side of the argument at the Cambridge Union debate. Then, it was up to the two human-led teams to elaborate on these points and create rebuttals.
Cambridge Union hosted W Churchill, the Dalai Lama, and Stephen Hawking. On Thu #IBM#ProjectDebater was added to the list. And I was there to give the intro. Best story is by @SciTech_Cathttps://t.co/MUr5gUIk1i#NeilLawrence#HarishNatarajan#SylvieDelacroix#SharmilaParmanandpic.twitter.com/NerWJGBjlR— Noam Slonim (@noamslonim) November 23, 2019
How would this technology be useful in the 'real world'?
The engineer who leads the Project Debater project at IBM, Noam Slonim, believes that this technology could be used by companies when they need to understand what their customers think of their products. Alternatively, it could be used to find out what employees think of a particular policy.
Furthermore, Slonim believes this tool could also be used by governments looking to better understand what their citizens think.
To Slonim, this is a way of showcasing how AI could work alongside human beings in the future, assisting in improving their working situations, and not competing against them.
The software remains imperfect, however, as Slonim noted that during the Cambridge debate the AI made some mistakes by using the same arguments twice, or by contradicting itself.
In the end, with a 52.1% majority, the Cambridge Union voted in favor of the proposition that AI would cause more harm than good.
You can watch the full debate here: