Computers typically have a vast amount of information at their disposal but IBM’s was unable to persuade a live audience of its viewpoint.

IBM’s debating computer was tasked with taking on Harish Natarajan, a grand finalist in 2016’s World Debating Championships.

The live debate was broadcasted live on Monday and held in San Francisco’s famous Yerba Buena Center. It was organised by a non-profit firm called Intelligence Squared US which specialise in hosting debates.

IBM’s system, known as ‘Project Debater’, has been in development since 2012. Project Debater uses newspaper and magazine articles to formulate its points rather than online sources like Wikipedia, which could explain why Natarajan got the better of it.

Debating AIs could provide various benefits. Politicians, for example, could use them for training ahead of live debates against opponents. They could also help with general decisions to determine whether an idea is flawed.

Monday’s debate focused on the question of whether preschool should be subsidised. Both the human and AI participant were not alerted to the subject until 15 minutes before.

The AI was on the side of subsidising preschool. It came up with strong arguments such as how subsidisation helps to break the poverty cycle. During the debate, it highlighted and drew information from various studies.

Natarajan argued that subsidies would not ensure that all children would be able to attend, while also highlighting that it consumes resources that could be used for other things.

Both sides gave a two-minute closing argument. While awaiting the final verdict, Natarajan noted how combining Project Debater’s skills with a human “could be incredibly powerful.”

Prior to the debate, the audience was asked how they felt about the issue. 79 percent said preschool should be subsidised, while 13 percent disagreed.

The winner is decided by which side gained the most points.

Following the debate, 62 percent of the crowd agreed and 30 percent disagreed. By converting 17 percent of the crowd, Natarajan was crowned the winner.

You can watch the full debate here.

deepgeniusai.com/">AI & Big Data Expo events with upcoming shows in Silicon Valley, London, and Amsterdam to learn more. Co-located with the IoT Tech Expo, , & .

The post Human beats IBM’s AI-powered debate computer appeared first on AI News.

Project Debater began its training six years ago but it could only hold a debate with people two years ago. A demo was held for journalists at IBM’s offices in San Francisco where the AI debated issues relating to healthcare and the subsidisation of space exploration.

The company didn’t give its AI an easy time either – it was competing against Dan Zafrir, President of the International Debate Society in Israel, and 2016 national Israel debate champion Noa Ovadia.

Debater opened its supporting argument for space exploration subsidisation with facts such as how it benefits humankind because it helps advance scientific discoveries and inspires young people to think beyond themselves.

Ovadia countered the argument saying there are better applications for government subsidies, including for scientific research here on Earth.

The AI considered Ovadia’s argument before delivering a rebuttal that potential technological and economic benefits from space exploration outweigh other government spending.

According to IBM, aside from opening remarks and jokes, none of the debates were scripted.

Following the debate, a poll showed the majority of audience members thought Project Debater enriched their knowledge more than its human counterpart.

Politicians have gained somewhat of a reputation for using arguments without factual basis in debates to support their campaigns. Short of their opponent being prepared to counter it, the argument can be seen as true. It’s easy to see where an AI like Project Debater could one day act as a mediator in these circumstances to ensure points are not misleading.

If nothing else, Project Debater would be great for training for any form of public speaking to ensure the arguments are factual and prepare for any potential rebukes to avoid embarrassment.

Heck, it could even be used privately to challenge some of our own views following potential exposure to politicians’ lies, “fake news”, and propaganda bots.

What are your thoughts on IBM’s Project Debater AI?

The post IBM’s AI will debate until you learn your mistakes, feeble human appeared first on AI News.