What can we learn from a debate between man and AI?
At IBM Think 2019, Project Debater, the “first AI system that can debate humans on complex topics” took on a champion debater, Harish Natarajan, on the topic, “We should subsidise preschool”. Project Debater argued for preschool subsidies while Harish argued against them.
Tapping into knowledge from 10 billion sentences, taken from newspapers and journals, Project Debater opened with an impressive four-minute speech, constructing reasons to support preschool education. While Harish was blown away by Project Debater’s ability to find relevant information for her argument and explain it in a straightforward, simplified way for the audience, she fell short on the human aspects of the debate, from comprehending the subtler arguments of an opponent to the ability to use emotion and intonation of voice to persuade. She lost the debate after the audience swung by 17% in Harish’s favour.
It would be easy to write off artificial intelligence as a novel, experimental phase in modern civilisation that isn’t yet relevant or useful in practical, commercial ways. But let’s take a look at predictive analytics, the branch of analytics that helps make decisions after sifting and sorting through thousands of choices. Predictive analytics solutions like IBM’s Decision Optimization enable organisations from commerce to healthcare make better decisions and achieve business goals by solving complex optimisation problems.
A business could say “people know best”, we don’t need predictive analytics tools. But what if they have to consider more than a thousand individual choices to make the best possible business decision? Decision Optimization considers these choices in seconds, rather than several days if processed by humans alone. In this case, AI is used to empower, not replace the people making decisions, by helping them plan better, faster, and shifting through more what-if analyses.
A business could also say that they don’t have all the data required to work with a solution like Decision Optimization. But if set up with good data governance, curation, and security, Decision Optimization isn’t just a set of unused tools. It becomes a real solution provider.
Artificial intelligence may never outpace us in its ability to make decisions or persuade through rhetoric and emotion, but it can certainly empower and complement us, and if we can understand and implement AI effectively into our organisations, it can give us a huge competitive edge over those that stay away from AI.
Need help with TM1?
We're here for you