White paper lays out Google position on government involvement in AI
Google has laid out its position on government regulation of artificial intelligence in a white paper, suggesting a limited approach to legislation, and calling for government involvement in five specific areas.
Perspectives on Issues in AI Governance (PDF) is a 34-page document which calls for governments to work with civil society and industry stakeholders on explainability standards, fairness appraisal, safety considerations, human-AI collaboration, and liability frameworks. The white paper also notes some of the trade-offs inherent to the discussion, such as between explainability and safeguarding security.
In the section on safety considerations, Google suggests that like electrical products in Europe are required to demonstrate their safety through CE certification, the same could be done for AI.
“For example, biometric recognition technology in smart lock systems could be tested against a representative, randomized dataset to ensure they exceed pre-set accuracy standards, before being certified safe for use.”
Google Global Policy Lead for Emerging Technologies Charina Chou, who co-authored the report, tells Wired that the motivation is to respond to government questions about what practical steps they can take to ensure responsible AI use. She cautions that further dialogue is needed before drafting new laws.
“At this time, it’s not necessarily super obvious what things should be regulated and what shouldn’t,” Choi says. “The aim of this paper is to really think about: What are the types of questions that policymakers need to answer and [decisions] we as a society have to make?”
Internet Institute researcher Sandra Wachter told Wired that high-level abstract claims such as that AI should be fair are of limited value at this point.
“I think it’s a good initial list. Where I’d say there is still a gap is how to govern those things,” she says.
Google’s suggestions are more abstract but also much broader than Microsoft’s recent proposals about regulation of facial recognition, and would apply to other biometrics as well. Lawmakers meanwhile continue to move ahead with proposed restrictions on biometrics and other AI technologies at various levels.