The big names in tech want the government to expand federal investment in artificial intelligence research and development. That is apparent in comments they have submitted in response to a recent request for information issued by the White House Office of Science and Technology Policy (OSTP).
More than 50 corporations, nonprofits, and researchers provided comments to a request for information to update the National Artificial Intelligence Research and Development Strategic Plan. OSTP published an abridged version of the responses in April.
The plan is a legacy of the Obama administration, which sought to “help the United States capitalize on the full potential of AI” for economic prosperity and social progress. It then received an update during the Trump administration with the most notable additions being the identification of eight strategic priorities for AI research and the National AI Initiative Act of 2020.
Comments were formulated around the eight following strategies from the 2019 version of the plan:
Strategy 1: Make long-term investments in AI research.
Strategy 2: Develop effective methods for human-AI collaboration.
Strategy 3: Understand and address the ethical, legal, and societal implications of AI.
Strategy 4: Ensure the safety and security of AI systems.
Strategy 5: Develop shared public datasets and environments for AI training and testing.
Strategy 6: Measure and evaluate AI technologies through standards and benchmarks.
Strategy 7: Better understand the national AI R&D workforce needs.
Strategy 8: Expand Public-Private Partnerships to accelerate advances in AI.
Receiving the most focus were Strategies 1, 4, and 8, although Strategy 3 also received considerable attention linked to suggestions of a new “Strategy 9” concerning environmental sustainability and international cooperation for AI research.
In particular, Strategy 1 and 8 were perceived as “critical” for sustained US leadership and innovation in AI technology:
“Breakthroughs in AI R&D now occur far too quickly for national strategic plans to be updated on a three-year cadence,” the AI safety and research company Anthropic wrote in its response, and without sufficient financial and temporal dedication to AI technology “the federal government risks missing critical research priorities and areas for additional investment.”
The Association for the Advancement of Artificial Intelligence pointed out that investment is not just limited to “larger data sets with more cloud computing resources,” but also should be concerned with creating “an environment that will enable such advances in order to support and reach the next level of AI.”
Respondents agreed there need to be more incentives to advance public-private partnerships, as such efforts are closely linked to Strategies 4, 5, and 7 (data privacy, public data sets, and the needs of the AI R&D workforce). Google further proposed that OSTP make a standalone pillar increasing AI talent in government—scouted from the private sector and otherwise.
These proposed improvements to AI research investments were seen as the basis for extensive research in environmental impact and algorithmic bias, areas respondents pointed to as primary concerns regarding the social impact of AI research (Strategy 3).
The Center for AI and Digital Policy reported that “AI-enabled systems require exponentially rising computing power…[and] substantial energy consumption, generating a huge carbon footprint and upending the green effects of digitalization.” To this purpose, the Center argued there needs to be greater interdisciplinary engagement between AI technologies and “environmental science, geology, …, astrobiology, etc.”
Bias issues are such that AI may “not work as well for one demographic group as it does for another.” Microsoft’s comments distinguished between “allocation harms, where a system may allocate resources or opportunities” and “representation harms, where an AI system may describe, depict, or otherwise represent people, cultures, or society…in demeaning [ways].”
Other notable suggestions included refining the plan’s visual presentation for better communication to the public, establishing new oversight organizations for increased transparency, and applying AI to drug discovery.
Respondents included Amazon Web Services, Microsoft, Google; researchers from Massachusetts Institute of Technology, Harvard University, and Cornell University; American Psychological Association and World Privacy Forum.