THANK YOU FOR SUBSCRIBING
Government CIO Outlook | Wednesday, January 15, 2020
OSTP has released ten principles in a draft memo to the heads of executive agencies that should govern the development and use of artificial intelligence (AI) technologies in the private sector.
FREMONT, CA: As part of the proposed set of the U.S AI regulatory principles, the White House Office of Science and Technology Policy (OSTP) urges the Federal AI regulators to limit regulatory overreach of the technology and its extended applications by the private sector. All the ten principles released in a draft memo to the heads of executive agencies should govern the development and utilization of artificial intelligence (AI) technologies in the private sector. According to the U.S. chief technology officer, Michael Kratsios, by building upon this Administration’s record of leadership in artificial intelligence, the proposed U.S. AI regulatory principles will set the Nation on a track of continued AI innovation and discovery.
Stay ahead of the industry with exclusive feature stories on the top companies, expert insights and the latest news delivered straight to your inbox. Subscribe today.
Here are the White House’s three primary goals for AI underlying the principles:
• Limit Regulatory Overreach: The regulators must conduct a risk assessment and cost-benefit analyses before taking any regulatory action on AI, thereby, establishing flexible frameworks rather than one-size-fits-all regulation.
• Ensure Public Engagement: Scientific evidence and feedback from the American public, the academic community, industry leaders, non-profits, and civil society is necessary to make technical and policy decisions.
• Promote Trustworthy AI: For deciding AI-related regulatory action, regulators should consider fairness, openness, non-discrimination, safety, transparency, and security.
Here are the ten OSTP principles:
• Public trust in AI- AI poses a risk to privacy, autonomy, individual rights, and civil liberties. AI’s continued adoption and acceptance will depend increasingly on public trust and validation. Therefore, the government’s approach to AI should promote robust, reliable, and trustworthy applications.
• Public Participation- At instances where AI uses information about individuals, there is public participation, and thus, will improve agency accountability and regulatory outcomes while increasing public trust and confidence.
• Scientific Integrity and Information Quality- Agencies should hold information either produced by the government or acquired by the government from third parties. It is because it is merely to have a clear and substantial influence on crucial public policy or private sector decisions, along with those made by consumers to a high standard of quality, transparency, and compliance.
• Risk Assessment and Management- Not all foreseeable risks should be mitigated. Moreover, a risk-based approach should be followed to distinguish acceptable and unacceptable risks or risks that result in severe loss than benefits.
• Benefits and Costs- Agencies should abide by laws and carefully consider the full societal costs, benefits, and distributional effects before taking into account the regulations related to the development and deployment of AI applications.
• Flexibility - Rigid, design-based regulations that prescribe the technical specifications of AI applications will be impractical and ineffective often, given the expected pace with which AI will evolve and the resulting need for agencies to respond to new information and evidence.
• Fairness and Non-Discrimination- In some cases, AI may introduce real-world bias results in discriminatory outcomes or decisions that chip away public trust and confidence in AI.
• Disclosure and Transparency- Besides improving the rulemaking process, transparency and disclosure can increase public trust and confidence in AI applications. In some instances, such disclosure may require identifying when AI is in use, for example, if suitable for addressing questions about how the application affects human end users.
• Safety and Security- Agencies should support the development of AI systems that are safe, secure, and operate as intended while encouraging the consideration of safety and security issues throughout the AI design, development, deployment, and operation process.
• Interagency Coordination- Agencies should work together to share experiences and to make sure consistency and predictability of AI-associated policies that advance American innovation and growth in AI, while adequately protecting privacy, civil liberties, and American values and allowing for the sector- and application-specific approaches when appropriate.
More in News