insight
May 22, 2023
A series of frameworks issued by various federal agencies that apply existing federal and state laws despite media headlines suggesting that there are no laws governing artificial intelligence (AI) in the United States also apply. In addition, AI-specific laws exist in many states, and additional states are considering laws addressing various uses of AI-enabled technology. As the AI regulatory landscape evolves in the United States, increasing scrutiny by lawmakers and regulators continues to be a common theme.
Federal regulatory trends
Since the last update, there has been significant federal activity related to AI regulation, led by the Federal Trade Commission (FTC). In February 2023, the FTC issued guidance on the applicability of enforcement powers under Section 5 of the FTC Act to address allegations of unfair or deceptive AI advertising.
Specifically, the FTC warned that it will focus on whether marketers make false or unsubstantiated claims about AI-powered products. For example, exaggerating the technical capabilities of the product, unproven claims that AI-equipped products are superior to non-AI-equipped products, or whether the product uses AI.
Under the FTC’s guidance, companies that release AI-enabled products or services should consider whether those products or services contain inherent biases that could make them unsuitable for particular uses. also warns that you should fully understand the risks and limitations associated with such products or services.
Additionally, on April 25, 2023, officials of the FTC and three other federal agencies (U.S. Department of Justice, Office of Civil Rights, Consumer Financial Protection Administration, U.S. Equal Employment Opportunity Commission) issued a joint statement pledging to: announced. While emerging automated systems, sometimes marketed as “artificial intelligence” or “AI,” are becoming more and more common in our daily lives, impacting civil rights, fair competition, and consumer protection, Supports America’s commitment to core principles of equality, justice, equal opportunity, and equal opportunity. ”
However, the FTC is not the only federal agency active in the AI regulatory space. On January 26, 2023, the National Institute of Standards and Technology (NIST) announced the AI Released Risk Management Framework version 1.0. AI system. Divided into two parts, the first assesses what a trustworthy AI system looks like, and the second details four categories of capabilities to address risks in AI systems. To do.
Despite its voluntary nature, the NTIA’s guidance could become industry best practice similar to what happened with the Framework for Improving Critical Infrastructure Cybersecurity. This framework is widely used not only by the majority of US companies, but also by many non-US companies. It is used as a model in the United States as well as in several other countries. NIST plans to release updated version 2.0 of its AI risk management framework in spring 2023.
State-level AI laws and regulations
In addition to the moves at the federal level, state legislature submissions of AI-enabled bills increased 46% between 2021 and 2022, and state special committees looking at the need for AI-centric regulation. Committee.
State legislative efforts address a myriad of topics, including predictive policing technology. Use of facial recognition technology by police. Consumer Focused Rights. Employment-related issues. Implementation of automated decision-making by the financial and insurance industries. and medical related issues.
Below is a summary of current AI regulations in each US state.
Legislation to support AI:
Laws under consideration to address AI:
Key considerations when considering and implementing AI into your business
- Identify AI applications in your operations. Companies should use a comprehensive set of operational decisions and processes to determine what operational decisions and processes they use to make decisions, where some of their decision-making processes rely on automated processes. A survey needs to be conducted.
- Conduct a documented risk assessment. Once identified, companies should assess the risks and mitigation potential of current and future AI tools. Firms analyze the costs and benefits associated with the introduction of new technologies, determine the appropriate technical, contractual and organizational risks associated with such technologies, and implement relevant controls within appropriate business units. You should consider establishing an internal mechanism.
- Integrate structural compliance measures at the foundational level across your organization. Companies should draft or adapt policies governing the use of AI applications within their organizations. Policies should cover common themes such as transparency, accountability and fairness, including dataset integrity, accuracy, foreseeable risk and social impact. Your policy should specifically address how your organization evaluates the implementation of AI-related technologies to ensure that they do not create unequal and unfair outcomes.
- Proactively designate responsibility for accountability and governance. Consider appointing an individual to develop, maintain, and enforce AI-related policies that function similarly to a chief privacy officer, depending on how pervasive or critical AI technologies are to the enterprise please.
- Adopt existing compliance/risk management programs where applicable. An organization’s approach to AI may fall within an organization’s existing risk management practices. Depending on the prevalence of AI within your industry and organization, consider aligning your governance policies with different business units to raise awareness and embed risk mitigation into your day-to-day operations.
For more information on US federal and state efforts to regulate AI and machine learning systems, see the presentation “Artificial Intelligence Regulation in the US”, part of the company’s technology marathon web series.
