IBM Research has built an AI system that can analyze 300 million articles, papers or records on a given topic and construct a persuasive speech about it. It would take a human—reading twenty-four hours a day—about 2,000 years to get through the same material.
IBM Project Debater does it in 10 minutes.
Rest easy. It won’t replace humans for reasons I’ll explain below. It will, however, have profound implications for how we make decisions to solve complex challenges.
Project Debater is the latest AI technology in IBM’s series of grand challenges. First, “Deep Blue” took on chess champion Garry Kasparov in 1997. In 2011, IBM Watson beat a human champion on the game show, Jeopardy! Now, Project Debater moves us closer to mastering human language, a key objective in AI research.
This week I spoke with Talia Gershon, IBM’s director of research strategy. She specializes in bringing AI technology and quantum computing from the research labs to commercial applications. According to Gershon, Project Debater proves an AI system can do far more than answer simple questions. “It is a demonstration of AI’s ability to understand language well enough to construct and represent a really persuasive point of view about a complex topic.”
IBM research scientists with expertise in computational argumentation (the science of analyzing text to create a system with logical reasoning) have been working on the AI debate system for six years. In 2018, IBM showed off Project Debater’s ability to take a statement it had never seen before—“We should subsidize space exploration”—and deliver a four-minute argument supporting the statement with facts.
Next month, the system will take on a world champion human debater in front of a live audience. They will both be handed the same topic and given ten minutes to form a speech about it. In addition to presenting one side of the argument, the AI system will have to listen to the human debater’s speech, understand it, and form a rebuttal. The researchers assure me Project Debater is up to the challenge.
Bringing AI to the Art of Debate. At CES 2019, IBM unveiled the next step in its AI debate technology. It’s called Project Debater–Speech by Crowd. On each day of the conference IBM introduced a statement about a topic and invited people to the booth (or online) to write a short sentence explaining their point of view.
The first statement was “Gambling should be banned.” Project Debater gathered all of the opinions, analyzed them, and returned two speeches—one that supports the statement and the second that opposes it. The system determined which opinions were the most salient and eliminated redundancies—recognizing which opinions it had already covered.
You can see both speeches here.
The AI-generated speeches are remarkably convincing. The system began its speech with a short summary:
“Greeting and welcome all. You are about to hear a speech supporting the idea that Gambling should be banned…”
The 332-word speech arguing that gambling should be banned offered three reasons (with evidence) to support its case: “Gambling is addictive,” “facilitates criminal activity,” and “has ruined many individuals and families.” The second speech—arguing that that gambling should not be banned—also provided three reasons.
Regular readers of my column know that “the rule of three” is a fundamental component of persuasion. Overloading a listener with too much information at any one time makes it difficult for humans to process the content. Project Debater already knows it.
Project Debater marks a major milestone toward understanding language. The AI system can complement human decision-making by bringing in facts and evidence in a persuasive, logical structure. By understanding people’s opinions on different topics, politicians, public servants and business leaders can get a better understanding of what people think about a policy or corporate decision–and why they think the way they do.
AI Systems Lack One Critical Human Quality. People understandably grow concerned when they hear about an AI system getting better at something we consider a uniquely human quality. Don’t fret. While Project Debater can synthesize human arguments into a reasonably coherent speech; it does not have feelings one way or the other.
Neuroscientists like Antonio Damasio have discovered that, without emotions, humans would be incapable of making even the smallest decision. Without emotion, “we wouldn’t have music, art, religion, science, technology, economics, politics, justice, or moral philosophy,” says Damasio.
After Garry Kasparov lost a chess match to an IBM machine in 1997, he said he felt “unsettled.” Today he says that humans and machines can work together to advance the world and to make better decisions. In a TED Talk, Kasparov said, “Machines have calculations. We have understanding. Machines have instructions. We have purpose. Machines have objectivity. We have passion…There’s one thing only a human can do. That’s dream. So let us dream big.”
IBM researchers are dreaming big and their dreams are helping us achieve a far greater understanding of human language—an understanding that might help us make better decisions about complex topics that face our world.