Artificial Intelligence Policy: A Primer and Roadmap

Artificial Intelligence, Innovation and Economic Growth and Privacy and Security

Article Snapshot


M. Ryan Calo


Available at:


Artificial intelligence (AI) systems are being deployed by commercial firms, and by the policy and the military. AI raises difficult policy issues, including questions of fairness, safety, and employment loss.

Policy Relevance

New rules, institutions, and policies will be needed to address AI policy problems.

Main Points

  • AI systems allow machines and computers to approximate human or animal cognition, even if the machines cannot really think.
  • AI systems may be unfair or biased; policymakers should consider what laws or incentives should be used to ensure fairness.
  • Some robots and AI systems will be capable of using force or weapons, and policymakers should consider requiring meaningful human control of such systems.
  • The ability of AI systems to recognize patterns and analyze information will create new privacy problems; consumers do not understand the full consequences of sharing information.
  • Policymakers should consider proposals such as a universal basic income, to reduce the potential affect of job losses as AI systems are deployed.
  • New advisory bodies could help introduce AI expertise into government agencies, which tend to lack understanding of how new technologies work.
  • The owners of proprietary AI systems are reluctant to explain how these systems work and allow outsiders to examine them.
    • Transparency is important to ensuring that these systems operate fairly and safely.
    • Trade secret law, the Computer Fraud and Abuse Act, and the Digital Millennium Copyright Act can impede accountability.
  • Some computer scientists fear the AI systems will attack and destroy humanity; even if this is unlikely, we should pay some attention to this problem, but should not allow it to distract us from other AI policy issues.


Get The Article

Find the full article online

Search for Full Article