While programs such as Artificial Intelligence bots that can write poetry or develop art are capturing people’s interest, administrative agencies across the country are concerned about how similar technology including algorithms and automated decision making may affect employees and consumers alike. The Equal Employment Opportunity Commission (EEOC) to the New York City Department of Consumer and Worker Protection are issuing guidance and regulations about AI and related technologies.
The latest administrative body to join the fray is the Colorado Division of Insurance. At the start of February, the Division issued a draft of proposed regulations pertaining to algorithm and predictive model governance. The purpose of the regulation is to establish requirements for a life insurance company’s internal governance and risk management necessary to ensure that the company’s use of consumer data and information, as well as algorithms and predictive models, does not result in unfair discrimination. This is a similar concern voiced in much of the guidance and regulations around the country.
The Division of Insurance’s proposed regulations includes governance and risk management framework, documentation mandates, and reporting requirements. The regulations would require life insurers that use external consumer data and information sources (ECDIS) as well as algorithms and predictive models using ECDIS to establish a governance and risk management framework that ensures the ECDIS is credible, and its use does not result in unfair discrimination.
That framework includes components that are similar to what we are seeing in other efforts to regulate AI and related technologies. These include:
- Documenting governing principles aimed at transparency and accountability.
- Board of directors and senior management’s responsibility and accountability for strategy and use of ECDIS and the algorithms and predictive models using ECDIS
- Establishing written policies and processes for design, development, testing, deployment, use, and ongoing monitoring
- Maintaining a process for addressing consumer complaints and inquiries, one that provides sufficiently clear information to enable consumers to take meaningful action in response to adverse decisions.
Additionally, as with other regulations in this area, insurers will be required to document their use of ECDIS and algorithms and predictive models using ECDIS, and report to the Division of Insurance progress toward compliance with the applicable regulatory requirements.
The type of regulation proposed by the Division of Insurance is going to proliferate as algorithms and automated decision-making tools become more and more common. As such, businesses exploring these technologies should consider putting similar measures and principles in place – e.g., governance, documentation, accountability, notice, and responsibility – during the design, development, testing, deployment, use, and monitoring phases.