The race to responsible artificial intelligence

Chances are you’ve heard of artificial intelligence, which has found widespread adoption in our popular culture. Generally speaking, AI is a set of computer algorithms that can learn to make decisions without human intervention. Autocorrect on our smartphones? That’s AI, though I think we can all agree there is some room for improvement in that arena. It’s fun to ponder what technological wizardry AI will offer future generations. Where it becomes daunting, however, is when we consider its long-term global implications and the structure necessary to harness its development.

How will AI improve our lives next? That depends. Right now, we are in a race to build effective AI that can be developed and used here in America. But it’s not a sprint or a 5k, it’s more like a marathon relay race on an uncharted course. To win this race, we need a team and a plan.  What do we get if we win?  We get to use AI that is developed consistent with our American values, and we can set the “rules of the road” for its use. What happens if another team wins?  The AI-enabled technology we use will be developed offshore, threatening American jobs in this burgeoning field and leading to applications that skew from our norms. We may find ourselves using algorithms that decide who gets a loan based on their cellphone’s battery-charging history. That’s already happening in China, and while it may be standard practice there, it certainly feels like an invasion of privacy to me. The use of AI can be highly influenced by the values of the society that creates it.

{mosads}The good news is that in the U.S., we already have some of the best scientists and engineers in the world working feverishly to come up with better algorithms to power AI devices. Those of us who experienced computer translation in its infancy recall that it often produced hilarious results; these days it’s pretty accurate, thanks to American AI innovation.   

We’ve got a great team, but we don’t yet have a plan. To navigate this uncharted relay race, we need to be coordinated and strategic about who is doing what, and when, with a commitment of the necessary resources. This is where America currently lags. Our counterparts in China have announced their vision to use AI pervasively in their society, making extensive use of video recording and other personal data in public and not-so-public spaces, with Chinese researchers working overtime to put AI-enabled technology in the hands of the Chinese government and consumers. In the UK, they are strategically cultivating a workforce to create and use AI-enabled devices. Here in the U.S., we don’t have a current strategic plan for AI research and development to get us to where we want to be, and without such a plan we don’t have much hope to prevail.

The recently-announced executive order directing agencies to prioritize AI’s development is a testament to the need, but it doesn’t provide resources or a plan to accomplish any goals.

Toward that end, my office has been working for months now to learn about the needs of our American AI research community, interviewing a wide variety of stakeholders including high-level experts at our National Laboratories and agencies, teachers who are looking out for their students’ future careers, workers who are now using AI in their jobs, and experts from large and small private sector companies who provide cloud computing resources. We’ve listened to what they had to say and carefully crafted the Growing Artificial Intelligence Through Research (GrAITR) Act, which I will introduce in the House of Representatives soon. This will establish a strategic plan to invest $1.6 billion over 10 years in our nation’s AI research, development, and education.  This bill will ensure coordination of AI-related science and technology across federal government agencies. The National Institute of Standards and Technology will create standards to evaluate the algorithms behind AI-enabled technology, charting the course toward responsible AI. The Department of Energy, with its world-class National Laboratories, will build next-generation computing systems to help AI learn faster from ever-larger datasets, paving the way for public-private partnerships to translate new AI discoveries into products that Americans can use every day. The National Science Foundation (NSF) will fund multidisciplinary research centers to study the basic science of AI and its effects on our society. The NSF will also develop AI education programs at both large flagship universities and smaller regional institutions for students young and old at every career stage, including retirees. This part is very important, to make sure that everyone is prepared for our AI-enhanced future. 

I want the benefits of AI to be shared by all Americans, not just a privileged few. As AI becomes more integrated into our everyday life, many American workers and consumers will be assisted by it, freeing us from our more tedious tasks and allowing us to do what we do best – think and make decisions. 

Lipinski is a senior member of the Science Research and Technology Subcommittee.