The Six Key Elements of a National Machine Intelligence Strategy for the U.S.

March 2018

“Machine Intelligence (MI) systems have already begun to quietly pervade a growing share of businesses, governments, and individual lives around the world, and we are only just beginning to grasp the impacts that this technological revolution will have on our economy, our society, and our national security.”

This post is adapted from the CSIS report, A National Machine Intelligence Strategy for the United States. Read the full report here: cs.is/2CPJCcW

What is Machine Intelligence?

MI refers to machines’ ability to perform tasks that would normally require human intelligence.

A national strategy for MI will provide the United States a platform for creating the policies and investments that ensure MI’s progress is consistent with American goals, norms, and values.

This strategy should have two overarching goals.

A. To promote the safe and responsible development of MI technology

B. To maintain U.S. leadership in MI.

Creating a National Machine Intelligence Strategy

1. Invest in the continuing research and development of MI technology

• China dwarfs all other countries in the scale of its research ambition. Two Chinese cities alone have pledged a collective of $7 billion to MI projects. In comparison, total U.S. government R&D investment in MI was $1.1 billion in 2015, and the Trump administration’s proposed budget would cut the National Science Foundation’s MI research funding by 10 percent.

2. Develop a workforce for the MI age

• It is likely that new technological advances will create more than enough jobs to make up for those lost through automation, but current workers will confront obstacles as they struggle to transition to new careers. Promoting investment in continuing education and job retraining will be crucial to managing these disruptions.

3. Create a flexible and open data ecosystem in which MI can thrive

• MI systems are data hungry. They require massive datasets to identify unique patterns and insights. One thing policymakers can do is expand access to datasets owned and managed by the U.S. government. Opening these datasets up to the public could have an enormous impact on the creation of new MI applications.

4. Create thoughtful public policy to spur MI adoption and reduce barriers

• Policymakers should also pursue measures that help spur demand for MI technologies, growing the market for MI tools and giving room for new U.S. MI enterprises to expand. The federal government can support the adoption of automation and MI by firms by removing legal and regulatory obstacles and creating incentives for MI adoption.

5. Develop proactive strategies to manage MI’s risks

• A well-thought-out strategy to manage MI’s challenges can help build consumer trust in the technology and help us to avoid falling into a trap of reactionary policy making when accidents inevitably happen. At the same time, we must ensure that any rules and controls we put in place make sense for the technology and do not strangle the commercial MI industry.

6. Lead the development of MI technology and MI governance around the world through strategic MI partnerships

• We should establish research partnerships with other MI leaders to develop new MI technologies. Strengthening the United States’ ability to collaborate with technology leaders in these countries will significantly improve our capacity for innovation and help keep the United States on the cutting edge of new MI development.

This post is adapted from the CSIS report, A National Machine Intelligence Strategy for the United States. Read the full report here: cs.is/2CPJCcW