The audience of assembled executives at Fortune’s Most Powerful Women Summit in Laguna Niguel, Calif. on Tuesday were in near-unanimous agreement that artificial intelligence was poised to do more good than harm in the world.
Still, the experts speaking before them said they believe the promise of A.I. comes with important caveats.
“To be totally transparent, we are every day wrestling with how we deal with this thing,” said Aicha Evans, senior vice president and chief strategy officer of semiconductor giant Intel. Top of mind: safety, privacy, and ethical considerations—not to mention concerns around cultural sensitivity and bias.
Evans cited self-driving cars as one example. She described a mathematical model Intel developed called Responsibility-Sensitive Safety (RSS), which establishes clear safety standards for autonomous vehicles while making sure the technology could be commercially scalable. “We need to be very responsible,” she said.
Bridget van Kralingen, IBM’s senior vice president for global industries, platforms, and blockchain, shared a similar vision. A fully A.I.-enabled future must arrive with careful consideration, she said. “We are working industry by industry to get A.I. into their hands,” she says, ticking through a list that included call centers, farming, fraud detection, and oil and gas project management. But, van Kralingen cautioned: “A.I. is not a fairy dust thing you can sprinkle. For it to be useful and enduring, it needs to go into your core processes.”
It’s all moving very quickly, the executives agreed, so trust is key.
For example, IBM years ago devised a set of principles governing its A.I. work that are meant to reinforce trust, van Kralingen said.
The first: Make sure people understand that their data is their data and how and where it’s used. The second: Be explicit about how the A.I. model learns and makes decisions or recommendations. The third: Demonstrate a clear commitment to using A.I. for the good of other human beings, including building human skills.
Avoiding unintentional harm and minimizing bias is every professional’s responsibility, said Toni Reid, who leads the Alexa unit at Amazon. “Be thoughtful about who you’re bringing it to the table when you’re building these products,” she said.
But, Evans noted, humans are still the best computing machines around.
“Life is about figuring out purpose, connecting dots, making judgments,” she said. “Computers don’t know how to do that yet.”