Debating is to discuss and argue on something about a particular subject. Humans can be good at this, and so should computers.
Tech giant IBM thinks that computers should also learn how to debate and argue like human beings, and one attempt is using a project calls ‘Project Debater – Speech by Crowd,’ which IBM describes as the "new and experimental cloud-based AI platform for crowdsourcing decision support."
Here, the company hopes to eventually unveil a system capable of engaging humans in unbiased debate.
But to make this possible, IBM needs the AI to learn as much as possible by debating and argue with people.
Since debating is something unique to humans in which we can all do well, IBM puts this Project Debater forward to solicit "arguments for and against a specific topic from as many humans as possible, and then uses them to create debate speeches.
The idea is to crowdsource people's opinions and combining them with AI‘s superhuman ability to crunch data and come up with patterns the AI can correlate for future problem solving.
The platform is also designed to help people practice their reasoning, develop well-informed arguments, and reach reliable conclusions.
According to IBM's Project Debater website page:
"Project Debater - Speech by Crowd uses AI to collect free-text arguments from large audiences on debatable topics and automatically construct persuasive viewpoints to support or contest the topic. After the AI analyzes the arguments for their polarity, strength and relevance to the topic, the results are displayed on this indicator scale."
IBM Research has been working on the Project Debater for at least seven years.
Led by Noam Slonim and Ranit Aharonov from IBM Research Haifa, the AI could be used in a variety of applications in many use cases. This includes business, education, law, and government, according to Aya Soffer, vice president of AI Technology at IBM Research.
In its quest to learn how to debate better, IBM collects text arguments from a wide audience, and uses the Project Debater AI to analyze the validity and strength of those arguments, to then organize the key points, and edit together an argument of its own.
Speech by Crowd "will be the first time we test [Project Debater] with the public," Soffer said.
"With that said, we envision the technology as an AI service in the IBM Cloud. … We are interested in speaking with clients about their use cases, which we could potentially explore together further."
The goal of this project is for the AI to remove bias in decision-making, and improve machine-human communication.
But the project is a work in progress and there are unforeseen possibilities.
In a blog post announcing the project’s public debut, researcher Noam Slonim and Ranit Aharonov said that:
The applications for Speech By Crowd are enormous. If successful, it could be used for every topic, and the outcome should have no bias, opinion or any agenda.
While it seems to be a good idea for IBM, but the company has to deal with the fact that humans do't argue fairly.
How AI works is by looking for patterns and learns from them, to then use that knowledge to solve future problems.
The machines here, relies on rational thought and logic debated by us humans.
If IBM has found a way to mitigate some issues that were present in AIs like Microsoft Tay or Microsoft Zo for example, then perhaps making AI smarter with debates is indeed the way to go.