What the US executive order says about the future of AI

The US signals intent to join other world players with broad AI regulation.

In the news

Last week, US President Joe Biden issued an executive order that, the administration ambitiously claims, “establishes new standards for AI safety and security, protects Americans’ privacy, advances equity and civil rights, stands up for consumers and workers, promotes innovation and competition, advances American leadership around the world, and more.”

The move was welcomed by many; if nothing else, the executive order demonstrates that the federal government grasps the potential of artificial intelligence—which, thanks largely to OpenAI’s ChatGPT and similar generative AI products, has broken out of the tech ghetto to become a consumer phenomenon.

Biden also called on Congress to emulate the European Union and China by passing new, sweeping federal data privacy legislation. Such legislation exists today but on a piecemeal or state-by-state basis. Moreover, later in the week, the US joined the EU, China and other nations in signing a document dubbed the Bletchley declaration that warns of AI’s potentially “catastrophic risks.”

Under the order, an unspecified agency would have oversight over the safety tests companies use to evaluate conversational bots (like ChatGPT). The National Institute for Standards and Technology would take the lead in developing standards and tests “to help ensure that AI systems are safe, secure and trustworthy.” A host of additional provisions would target everything from safeguarding national security to labeling AI-generated content to advancing sustainable development.

We wondered what message the Biden order sends to developers, business leaders, consumers and the world.

The Cognizant take

According to Tahir Latif, Global Practice Lead for Data Privacy & Responsible AI at Cognizant, perhaps the most important aspect of the executive order is its very existence. “It signals intent and increased federal attention to reducing the risks around AI,” he says.

Like others, Latif has concerns that the US system of government will make it difficult for the nation to institute AI guardrails, just as it has thus far been unable to enact data privacy legislation on the scale of the EU’s General Data Protection Regulation or China’s Personal Information Protection Law.

As to substance, Latif says, the order is about accountability. “Previously, you could design, train and deploy AI without real risk assessment. The order places greater scrutiny on data practices to ensure organizations consider these risks.” With accountability comes transparency, he adds. “You have to tell your audience whose data is being used, and how it’s being both procured and used.”

For business leaders, Latif says, the executive order serves as a framework. “What you need to do in order to responsibly develop and implement AI is being more clearly defined. As to how a business accomplishes this, he adds, dozens of frameworks already exist, including those from the OECD, the EU, the IEEE and IBM.

By their nature, presidential executive orders tend to be sweeping, somewhat vague and overly optimistic. While that may indeed be the case here, Latif says the US has made a good start on coming to grips with one of the most important technological shifts of our lifetime.