Man Versus Machine Index Fight Hinges on People Trusting Robots

(Bloomberg) -- Would you let a robot pick your investment portfolio?

Deutsche Boerse AG’s Stoxx Ltd. indexing business is set to find out with its two latest market gauges. The first, the Stoxx Global Artificial Intelligence Index, will use human-led research to select companies involved with artificial intelligence technology. While the second, the Stoxx AI Global Artificial Intelligence Index, will deploy machine learning to pick those companies.

If you’re reaching for your wallet, hold up. Neither index has been licensed to a fund provider yet. But Matteo Andreetto, the chief executive of Stoxx, is optimistic that ETFs and mutual funds in the U.S. or Europe will latch on before long.

“It’s man against machine,” he said from the sidelines of the “Inside ETFs” conference in Florida this week. “Are [investors] comfortable with the humans? Or do they trust the machine and the algos more?”

The human-created index is comprised of companies that generate more than 50 percent of their revenue from AI-related business. The second index, however, is “AI squared,” as Andreetto puts it, with machines picking AI stocks.

Here’s how it works: Using natural language learning, a computer sifts through patents to identify those related to AI. An algorithm then calculates how significant the AI operation is within the company’s overall intellectual property and also how important it is compared to other AI patents. Stoxx teamed up with Yewno Inc., a Silicon Valley-based AI company, to develop this gauge of roughly 200 companies, which includes Apple Inc., Bank of America Corp. and Facebook Inc.

So far, the humans have shown an edge in generating pure returns -- 6.1 percent to 5.4 percent this year -- while the robots are proving better at limiting volatility. But it’s still early. There’s plenty of time for the rise of the machines.

To contact the reporters on this story: Rachel Evans in New York at revans43@bloomberg.net, Kailey Leinz in New York at kleinz1@bloomberg.net.

To contact the editors responsible for this story: Jeremy Herron at jherron8@bloomberg.net, Eric J. Weiner, Andrew Dunn

©2018 Bloomberg L.P.