Have Your Data and Use It Too: A Federal Initiative for Protecting Privacy while Advancing AI
Summary
The Biden-Harris Administration should aim to make the United States a world leader in privacy-preserving machine learning (PPML), a collection of new artificial intelligence (AI) techniques capable of providing the benefits of machine learning while minimizing data-privacy concerns. By some estimates, improvements to the speed, accuracy, and scale of AI could augment global GDP by 14%, or $15.7 trillion, by 2030. Yet Americans fear that expansion of AI will have moderate to severe negative consequences. They are particularly concerned about the privacy implications of how companies and agencies use personal data to generate new developments. To assuage these concerns, this proposal recommends targeted initiatives for the Biden-Harris Administration to bring PPML techniques to maturity, including
- Investing in PPML research and development.
- Identifying compelling opportunities to apply PPML techniques at the federal level.
- Creating frameworks and technical standards to facilitate wider deployment of PPML techniques.
In an industry with such high fixed costs, the Chinese state’s subsidization gives such firms a great advantage and imperils U.S. competitiveness and national security. To curtail Chinese legacy chip dominance, the United States should weaponize its monopoly on electronic design automation software.
The technical advances fueled by leading-edge nodes are vital to our long-term competitiveness, but they too rely on legacy devices.
To tackle AI risks in grant spending, grant-making agencies should adopt trustworthy AI practices in their grant competitions and start enforcing them against reckless grantees.
As people become less able to distinguish between what is real and what is fake, it has become easier than ever to be misled by synthetic content, whether by accident or with malicious intent. This makes advancing alternative countermeasures, such as technical solutions, more vital than ever before.