Artificial intelligence (AI) is created in the image of humans. For thousands of years, humans have perfected the art of debating and arguing in all human societies and are considered fundamental capabilities of human intelligence. 

When AI develops argumentative capabilities, it is a monumental step towards further advancement of the technology. IBM has created a programme that is good at debating against humans. While it doesn't beat humans yet, it is only a matter of time that it does. 

The autonomous debating algorithm, Project Debater, is being developed by scientists at IBM for a few years. Now, it has developed the capabilities of arguing meaningfully with humans on a particular topic, taking a stand on a topic and arguing to make a strong point for its viewpoint. 

The new study was published in Nature in March. While it still has a long way to go to beat a human in argumentative logic, its public demonstration in 2019, showed off its abilities in reasoning. It argued against expert debating champion Harish Natarajan, on whether preschool should be subsidized for families. Here, it showed its capabilities to form complex arguments and present a case in support of the argument. This is a new frontier for AI as it has been used to processing and producing knowledge but so far, reasoning - an essential part of debating - was not a forte.

Project Debater accomplished this great due to argument mining. It entails that an AI will consume a huge amount of disparate information, searching for links between relevant sections. It has been fed over 400 million news articles, which enables it to design opening statements, rebuttals, and closing summations for at least 100 debate topics. 

"Debating represents a primarily cognitive activity of the human mind, requiring the simultaneous application of a wide arsenal of language understanding and language generation capabilities, many of which have only been partially studied from a computational perspective (as separate tasks), and certainly not in a holistic manner," the researchers, led by principal investigator Noam Slonim explain in their study.

"Therefore, an autonomous debating system seems to lie beyond the reach of previous language research endeavours," concluded Slonim. According to Slonim, this development will help improve human-AI interaction on a personal level. 

As this technology develops further, it will enable AI to become an inclusive part of the human decision-making process. We make decisions by weighing the pros and cons, experiences and knowledge. Project Debater, when developed further, can become a sounding board for a person seeking guidance, in the courtroom, etc. The technology will help improve virtual assistants too that would be able to hold conversations based on the queries, enabling them to be much more natural. 

Project Debater studies arguments and texts to build knowledge and arguments that means it amplifies a collective understanding which can help in greater decisions. This was demonstrated in an experiment, where IBM collected 3,500 opinions from the people of Lugani, Switzerland to hear their opinions of whether the city should fund autonomous vehicles. They used the Project Debater to assess these arguments for and against the proposal so that the local government could use the results to come to a balanced policy decision. 

“What we are doing touches on something fundamental to our lives,” he says. “We’re trying to tie language-understanding technologies together to help people make better decisions.”

Image from IBM Research