When it comes to taking a quick pulse of workers’ sentiment, surveys are still the most popular tool for soliciting feedback. Questions are straightforward and answers are in a multiple-choice format.
But what happens when HR leaders are confronted with workforce issues that are far more complex – even debatable? Those that can’t be answered with a simple “yes” or “no” and require insight.
This is where IBM is “bringing artificial intelligence into the art of debate” with a new cloud-based platform dubbed Project Debater – Speech by Crowd.
IBM unveiled the Speech by Crowd platform at CES 2019 to demonstrate how the AI can collect and analyze free-text arguments from the audience.
“We built a website and asked people to submit their arguments in 36 words or less, whether they were pro or con, and once they submitted them to the tool, the AI system would analyze them for their polarity,” explained Christopher Sciacca, communications manager for IBM Research.
The AI sifts through the raw data – phrased in natural human language and kept anonymous – then clusters the answers into categories to determine which side gained popular support.
The arguments help the AI formulate an overall stand. But while the AI is able to “debate” with an opponent using crowdsourced ideas, the AI is ultimately built as a decision support tool.
“Instead of having people vote, they can use Speech by Crowd to collectively get these arguments about why a new policy might be good or bad for the company.”
“We actually did this within IBM,” Sciacca said. “We used Speech by Crowd within our employee population and asked them about privacy and social media, and ‘Is social media good for society?’”