How A.I. Is Changing Virtually Every Business, from Commodities to Cosmetics
Companies are well aware that many A.I.-powered algorithms contain an inherent bias that could zero in on and discriminate against women, the elderly and certain ethnic groups.
Airlines are using artificial intelligence to improve punctuality. Credit card companies are relying on it to root out fraud. Consumer brands are using it to shore up their supply chain, and to ward off risk in the volatile commodities market where a single corner-the-market trade by a hedge fund or a Trump tweet could send raw materials prices spiraling.
Company after company on Wednesday at World Summit A.I. in Amsterdam shared their initial forays into deploying A.I. to address some of their more persistently vexing issues.
“We use these [A.I.] systems to examine as many as 100 external data sources—everything from Twitter to the bond markets—to determine future raw material prices, and the best time to buy,” said a chief data officer at one Fortune 100 company.
(As this statement and others were made under the so-called Chatham House rules in which the comments are made publicly, but in an unattributed fashion, Fortune is not naming the company nor executive who shared the story.)
A cosmetics brand told of how it used facial-recognition in a recent marketing campaign in which customers could upload a picture of themselves to get personalized beauty tips—for example, which kind of makeup would work best match their skin tone and personal style.
All in all, the executives remain bullish on the technology, but warn the use of A.I. introduces new organizational issues to confront.
For example, the executive at the cosmetics brand stressed the campaign worked only because the brand was transparent with the customer about which personal data it would ultimately use and store. “We have a legal, and, more importantly, a moral duty to be 100 percent transparent and accountable,” the executive said.
Throughout the day, tips on how to manage and protect customer privacy were shared. The matter has gotten only more urgent with the recent passage of the GDPR (General Data Protection Regulation) in Europe and the forthcoming equivalent in California, CCPA (California Consumer Privacy Act) where firms must be more transparent about how precisely they will use and store individual’s data, the lifeblood of many consumer-facing A.I. algorithms.
Another hot topic was that of bias. Companies are well aware that many A.I.-powered algorithms contain an inherent bias that could zero in on and discriminate against women, the elderly and certain ethnic groups.
As one executive put it, bias is most often observed when there’s a human in the loop. “Bias doesn’t seep into our data collection,” the executive said. “If it seeps in at all, it’s when humans begin to act on that data.”
More must-read stories from Fortune:
—How the man who nailed Madoff got GE wrong
—Most Powerful Women 2019: See who made the list
—Meet the women leading Netflix into the streaming wars
—Why the next recession may feel very different than 2008
—Old Navy is about to sail away from Gap Inc.—and into some choppy waters
Subscribe to Fortune’s The Loop newsletter for the latest on the revolutions in energy, technology, and sustainability.