“I don’t understand AI, and neither does anyone else.” This candid admission from AI expert Helen Toner during a recent TED talk highlighted a crucial challenge — even those building artificial intelligence systems struggle to fully comprehend how they work under the hood. Yet as Toner argues, this lack of clarity should not paralyze us from governing AI’s development and deployment.

Toner stressed that “AI is already all around us, so we can’t just sit around and wait for things to become clearer. We have to forge some kind of path forward anyway.” Inaction risks ceding control to major tech companies who may prioritize profits over societal impacts. Toner urges “don’t be intimidated” by the technology’s opacity or companies’ claims of specialized expertise required to regulate it.

Instead, she advocates policies focused on “adaptability, not certainty.” Since AI’s trajectory is difficult to predict precisely, Toner recommends “having plans in place for what to do in different scenarios.” Specific proposals include better measurement of AI capabilities, requiring companies to share system information for external auditing, and incident reporting systems to learn from real-world issues.

“If AI is progressing fast in dangerous directions, these policies will help us see that. And if everything is going smoothly, they’ll show us that too, and we can respond accordingly,” she said. Proactive yet flexible governance allows course correction as the technology evolves.

Ultimately, Toner reminds us that “we each have a voice in what happens…we are users, we’re workers, we’re citizens.” While complete foresight may elude us, pragmatic steps can steer AI’s influence productively by giving “as clear a picture as we can get of how the technology is changing, and then we can get in the arena and push for futures we actually want.”

Developing smart AI policies is undoubtedly challenging given the uncertainties involved. However, as Toner highlights, it is an essential endeavor to ensure this powerful technology benefits humanity as a whole rather than concentrating its impacts unhealthily. An open, adaptive approach to governance is key to responsibly harnessing AI’s immense potential.

Featured image: Credit: TED