IBM’s Project Debater, an argumentative AI-powered computer, lost a debate against Harish Natarajan, a grand finalist in 2016’s World Debating Championships. Both sides were informed of the topic only 15 minutes before the actual debate, but the human champion managed to persuade the audience and win.
Held at the Yerba Buena center, in San Francisco, the debate was organized by Intelligence Squared US, a non-profit specializing in hosting debates.
The Topic: Preschool Subsidies
The topic of the debate was whether or not preschool should be subsidized. Asked to argue for the proposition, Project Debater used compelling arguments such as how subsidies can help fight poverty, with information backed by various sources.
Natarajan argued that subsidies can consume important resources that could be used otherwise, and do not actually guarantee that the desired outcome is achieved – namely, that all children would be able to attend preschool.
The victory was awarded based on the percentage of the audience that each side managed to convert: 79% of the audience was initially in favor of subsidizing preschool, while only 13% were against. Following the debate, the percentage of those against – the view argued for by Natarajan – increased to 30%, thus making him the winner.
Not A Complete Loss
Though it lost the debate, the AI did manage to come ahead in terms of “enriching the knowledge” of the audience, with a majority of those present stating that they felt they learned more from the machine.
Natarajan also praised his opponent, stating that an AI-human collaboration can produce impressive results: “I think if you take some of those skills and add it to a human being who can use it in slightly more subtle ways, I think it could be incredibly powerful.”
In development since 2012, Project Debater has had previous experiences arguing with humans, with similarly impressive results.
According to Project Debater manager Ranit Aharonov, the purpose of the AI is not to actually win against humans, but rather to promote better decision making.
Aharonov says: “The vision around Project Debater is how do we develop the technology. It’s not a question of is AI going to be better at debating humans; that’s not really an interesting goal. The goal of this demonstration and of building this technology was to set [a goal] that’s far away and that enabled us to develop technologies to find relevant information within a mass of text, organise it and bring it to a position that’s digestible to humans, in order to drive better decision making more quickly.”
Potential applications of the AI are numerous, including political speechwriting and law.
The full debate can be found at https://twitter.com/IBM/status/1095125585058283520.
For more information, please visit https://www.artificialintelligence-news.com/2019/02/12/human-ibm-ai-debate-computer/ and https://www.itnews.com.au/news/ibms-project-debater-loses-live-debate-to-human-champion-519197.