The National Association of Insurance Commissioners (NAIC) is actively addressing the increasing use of artificial intelligence (AI) models in the insurance sector. According to a recent survey presented during a meeting of NAIC’s Big Data and AI working group, up to 40% of insurance companies are employing third-party AI models. To promote transparency and fairness in the utilization of these models, NAIC has outlined five core principles that guide the industry. Although these principles are not legally binding, they set expectations for insurance companies, AI stakeholders, and regulatory bodies, ensuring that AI models adhere to high ethical, fair, and safe standards.
The five core principles are as follows:
- Fair & Ethical: Acknowledging the risks associated with AI, this principle emphasizes that insurance companies must comply with regulations and implement controls to ensure that AI models are free from bias, prioritizing consumer interests.
- Accountable: Emphasizing transparency and accountability, this principle urges insurance companies to maintain comprehensive documentation of AI algorithms, tracking their evolution while adhering to jurisdiction-specific laws and regulations.
- Compliant: Recognizing the importance of compliance with both state-wide and federal laws, NAIC stresses that adherence to regulations is an ongoing commitment, requiring AI systems to consistently align with local, regional, and federal laws.
- Transparent: To build public confidence, AI stakeholders should commit to safeguarding proprietary algorithms’ confidentiality while providing easily understandable documentation disclosing the factors underlying predictions, recommendations, or decisions.
- Secure, Safe, and Robust: AI/ML algorithms used by insurance companies should be robust, secure, and safe throughout their lifecycle. Traceability is crucial, requiring detailed documentation and in-depth analysis in accordance with industry best practices and regulatory requirements.
However, challenges arise in implementing these principles due to the dynamic legal landscape and variations in state-specific laws and regulations. The ever-changing nature of laws and the diversity in regulatory requirements demand additional resources for insurers to navigate effectively.
To address these challenges, advanced solutions like Chiron Enterprise are proposed. Chiron Enterprise aims to assist insurers in achieving and maintaining regulatory compliance by capturing information, enforcing compliance through tracking metadata, and facilitating reporting on compliance. The platform allows for the translation of legal and regulatory requirements into mandatory controls, business processes, or subprocesses, ensuring comprehensive documentation and creating workflows that serve as audit trails.
Enforcing compliance through Chiron Enterprise is simplified, as the platform tracks the completeness of required metadata, establishing the model’s status and health throughout its lifecycle. Quality gates are implemented through deployment pipelines, ensuring that models can only progress to production if they meet compliance standards.
Reporting on compliance is facilitated by standardizing content elements across models and enabling users to select required content per state or legislation. Chiron Enterprise enables the generation of tailored documentation that serves as proof of compliance.
In conclusion, the complexity of ensuring compliance with evolving regulatory requirements, especially considering state-specific variations, necessitates the use of advanced technological tools such as Chiron Enterprise. These tools enable insurers to adapt to changing demands, navigate legal complexities, and demonstrate adherence to regulatory standards in the dynamic landscape of AI in the insurance sector. Read full article here.