Buried in the National Defense Authorization Act for Fiscal Year 2022 (NDAA), recently signed by President Joe Biden, are two of the most consequential pieces of artificial intelligence (A.I.) Legislation ever enacted into law: the Artificial intelligence Capabilities and Transparency (AICT) Act and the Artificial Intelligence for the Military (AIM) Act.
For the first time ever, Congress has signaled that the federal government is finally moving towards defining A.I. Ethics as a core requirement of the U.S. National strategy, while also asserting that traditional American values must be integrated into government and Department of Defense (DOD) A.I. Use cases.
While this legislation falls far short of the calls for regulation consistent with the European Union model and desired by many in the A.I. Ethics community, it plants the seeds of a thoughtful and inevitable A.I. Ethics regulatory regime.
Overall, the AICT and AIM Acts seek to accelerate the federal government and DOD’s ability to compete with the geopolitical reality of China’s (and, to a lesser extent, Russia’s) attempts to use artificial intelligence in ways that threaten the national security and economic interests of the United States.
Notably, the AICT defines A.I. Ethics as “the quantitative analysis of artificial intelligence systems to address matters relating to the effects of such systems on individuals and society, such as matters of fairness or the potential for discrimination.”
On its face, a definition of ethics that focuses on the quantitative aspects of A.I. Seems absurdly narrow. However, I believe it represents the seeds of a pathway that encourages modern A.I. Practitioners to harmonize algorithms with long-held ethical and moral principles.
It was German mathematician Gottfried Wilhelm Leibniz, a renaissance man well-versed in moral principles and widely regarded as an inspiration to the modern field of data science, who was the first to introduce the notion that mathematical algorithms have ethical consequences, and their impacts on society must be considered a priori.
While the bills in the 2022 NDAA are a starting point, operationalizing ethics into the A.I. Workflows and use cases of the federal government and the DOD requires more focused directives.
This legislation has put this task in the hands of the director of the National Science Foundation and its Artificial Intelligence Research Institutes, and there are three things the Director should do to ensure A.I. Development and deployment embeds ethics into every step of the process and complies with the new legislation:
- Create an A.I. Use case archive. A use case archive will allow visibility into every A.I. Use case in the federal government and DOD. Of course, access to this archive should be limited to those with the appropriate security clearances, but for Congress to effectively perform its oversight function and ensure implementation of its legislation, this is a necessary step.
- Harmonize existing A.I. Ethics vetting frameworks. The director should compile the various ethical vetting frameworks currently being deployed across the federal government, DOD, and private industry. For example, a promising set of comprehensive Responsible AI Guidelines was recently released by the Defense Innovation Unit at the DOD. Thus far, this is the most robust framework publicly released by the DOD to guide its contractors in the development and deployment of A.I. Use cases. The World Economic Forum A.I. Government Procurement Guidelines and EU Guidelines on Ethics in Artificial Intelligence are also worth reviewing.
- Develop a public communications strategy for A.I. Ethics. To get the American public and our allies around the world to trust that long-held American values are not being lost in the geopolitical competition, the government must make it abundantly clear how our development and deployment of A.I. Use cases is centered around our nation’s core values.
This communications strategy should be aggressive and transparent, including a quarterly or annual report of ethical frameworks being deployed in government A.I. Use cases, regular appearances by the Director of the National Science Foundation at the leading academic and industry conferences, and direct testimony from these parties at public congressional hearings.
Overall, this strategy should demonstrate how exactly the A.I. Ethics mandates of the AICT and AIM Acts are being implemented, as well as underscore the differences between the United States’ national A.I. Strategy and those of its geopolitical competitors.
Will Griffin is Chief Ethics Officer of Hypergiant, an enterprise A.I. Company based in Austin, Texas. Will is the recipient of the 2020 IEEE Award for Distinguished Ethical Practices and the creator of Hypergiant’s Top of Mind Ethics (TOME) framework which won the Communitas Award for Excellence in A.I. Ethics. He is a graduate of Dartmouth College and Harvard Law School.
More must-read commentary published by Coins2Day:
- We need a radical new approach to tackle scientific misinformation online
- The Great Resignation’s culture conundrum
- How the European gas market fuels Putin’s obsession with Ukraine
- Arianna Huffington: It’s time to replace work/life balance with “life-work integration”
- How George Washington defeated a pandemic to save the American Revolution