Accelerating AI Investments: Government Takes Steps to Lesson Regulatory Concerns
US Government views AI as vital to economic growth and global leadership and formally plans a minimized regulatory environment. This is great news for businesses since worldwide spending on Artificial Intelligence is expected to double in four years, reaching $110 Billion in 2024.
Takeaways
- The US supports and encourages the use and development of AI and officially recognizes that AI will improve competitiveness, fairness, and welfare.
- The government strongly supports a nonregulatory approach to ensure speedy progress.
- Guidelines balance AI innovation against potential risks and define government’s role.
- Investments in AI applications are unlikely to be constricted by regulatory actions and will, when voluntary shared, impact broader learning about important topics such as AI strengths and weaknesses, biases, risks, competitive impacts, privacy and personal decision-making, and national security implications.
Quick AI Timeline
- AI was popularized in the 1950s and has seen formal US governmental statements & actions accelerate rapidly only in recent history.
- The US is not alone in its quest for AI dominance. Foreign leader Vladimir Putin was quoted in 2017, “Whoever becomes the leader in this sphere [AI] will become the ruler of the world.” And, in that same year, China released their “New Generation AI Development Plan” worth US $150,000,000,000 ($150B) and stated intent to be the global leader in AI by 2030.
- Feb 2019 — Executive Order creates American AI Initiative — Includes five pillars: 1) Invest in AI R&D, 2) Unleash AI Resources, 3) Set AI Governance Standards, 4) Build the AI Workforce, and 5) International Engagement & Protecting our AI Advantage
- Nov 2020 — Memo to Heads of Executive Departments and Agencies — Guidance for Regulation of Artificial Intelligence Applications (summarized below)
- Dec 2020 — Executive Order focused on Trustworthy Use of AI by Federal Government
Guidance for Regulation of Artificial Intelligence Applications
The United States government is encouraging innovation and growth in artificial intelligence. Executive order 13859, Maintaining American Leadership in Artificial Intelligence, not only recognizes that AI is here to stay and galvanizes development and advancement, but the approach is designed to foster innovation and growth and engender trust. It aims to encourage businesses to move forward with developing and deploying AI without adding unnecessary barriers in the form of regulatory and nonregulatory actions. In fact, the supporting memorandum specifically states that agencies should forgo regulatory action when possible.
This is great news! The US government now sets the highest priority on AI innovation and growth and encourages regulatory agencies to follow the below guidelines:
Public Trust in AI
Public trust is necessary for the continued adoption and acceptance of AI. The role of the government is to encourage, with regulation or other actions, the following: AI does not compromise the decision-making abilities of individuals, privacy is paramount, and that AI be reliable, robust, and trustworthy.
Public Participation
Agencies are required to give the public ample opportunity to contribute and participate in the rulemaking process for AI.
Scientific Integrity and Information Quality
The government’s regulatory and non-regulatory approaches to AI applications should use scientific and technical information and processes. Utmost transparency is expected regarding strengths and weaknesses, biases, risks, competition impacts, privacy and personal decision-making, and national security implications.
Risk Assessment and Management
There are risks to AI and it is not necessary or possible to mitigate them all. There are trade-offs and a risk-based approach is encouraged to determine those risks that are acceptable and those that could cause unacceptable harm.
Benefits and Costs
Agencies should carefully consider societal costs, benefits, and distributional effects when considering regulations. The cost of doing nothing (not pursuing AI or too harshly prohibiting AI through regulations) should be actively considered.
Flexibility
It’s understood that rigid, design-based regulations are usually impractical and ineffective, especially considering the rapid pace of AI evolution. Regulations should not jeopardize flexibility.
Fairness and Non-Discrimination
AI has the potential to reduce discrimination caused by human subjectivity, but it can also introduce biases that produce unintended discriminatory outcomes. Regulatory approaches should consider whether AI reduces or increases discrimination.
Disclosure and Transparency
Disclosures must be written in a form that is easy for the general public to understand and to improve human decision-making.
Safety and Security
The development of AI systems that are safe, secure, and operate as intended should be promoted. Regulatory agencies must be mindful of any potential security and safety risks and vulnerabilities as well as the risks of potential malicious deployment.
Interagency Coordination
Agencies should coordinate with each other to ensure consistency and predictability of policies related to AI. Privacy, civil liberties, national security, and American values should be protected while allowing sector and application-specific approaches.
In Summary
The government is taking positive steps towards encouraging and democratizing the development of AI. Agencies are encouraged to lean towards nonregulatory actions to enable the “unleashing” of capabilities. Any regulations that are created should further the development and implementation of AI applications and engender public trust.
Further, the US government recognizes that development of AI is essential for America’s continued status as a global leader. This is great news for businesses as the government is firmly behind them in their efforts to implement AI applications in all parts of their organizations. If you’ve already invested, continue your work! If you haven’t, today is the best day to start.
