AI can see through your language
As some investors harness AI to learn and analyze their speech patterns and intonations, CEOs and other managers are coming under increasing scrutiny.
In late 2020, some IT executives downplayed the possibility of semiconductor chip shortages while discussing supply chain disruptions, according to Evan Schnidman, an expert on language patterns software. According to algorithmic analysis, their tone of voice revealed a high degree of uncertainty. The algorithmic analysis aims to find hidden clues in speech, preferably unscripted speech.
"We found that the tone of IT executives was at odds with the positive sentiment in their statements," says Mr. Schnidman. Schnidman advises two fintech companies on analytics. In the months since his comments, several companies, including Volkswagen and Ford, have warned that a severe shortage of chips has hit production. Then shares in the car and industrial companies fell. Mr. Schnidman argues that computer-driven quant funds would have been in a better position before the industry turmoil, by analyzing scores for the way managers speak rather than by reading the written word.
One example doesn't prove the accuracy of the presentation analysis, though, because we don't know whether executives were overly optimistic to begin with or genuinely changed their views as circumstances changed. But some investors see the technology, known as natural language processing (NLP), as a new tool that could give them an edge over rivals, according to 11 fund managers who are using or testing such systems.
Traditional financial data and corporate statements are now so heavily mined that they have little value, they say.
NLP is a branch of artificial intelligence where machine learning understands it through language and then turns it into a quantitative signal that quantitative funds factor into trading. The most ambitious software in this area is designed to analyze voice intonation, rhythm and emphasis, as well as phrases, while others attempt to analyze the text of speeches and interviews in increasingly sophisticated ways.
Slavi Marinov, head of machine learning at Man AHL, told Reuters NLP was "one of the main research areas" for the computer-driven fund. "These models take something very confusing and turn it into something that quantitative experts can easily understand," he says.
These AI systems can cost millions of dollars to develop and run, making it impossible for many investors and developers to invest in companies with deep pockets or niche markets. Some companies are still relatively experimental, and there are no public figures showing they make money. Funds interviewed declined to provide evidence that NLP could improve returns, citing commercial sensitivity.
However, some studies suggest that these technologies can improve performance when concentrated in smart places.