Artificial Intelligence Communications Services

The vendor is required to provide artificial intelligence (AI) would be used for administrative and non-emergency calls but could potentially expand to 911 calls.
- Provide a 45- minute demonstration of their AI technology.
- Generative AI is a type of artificial intelligence that can create new content or data based on existing data or inputs. 
- Generative AI include natural language interfaces to government services, image synthesis, and content creation. 
- Generative AI has the potential to enhance the efficiency, creativity, and innovation of government services and operations. 
- It must be deployed in a manner that mitigates risks to privacy, security, accuracy, and bias. 
- This policy aims to provide a framework for the responsible use of generative AI in government. 
- It outlines the principles, requirements, and best practices for using generative AI in a way that respects the county’s values, laws, and other standards.
- This policy applies to employees of the organization and to third parties including consultants, vendors, and contractors developing or providing generative AI tools to county.
- Principles
1. Fairness and responsible use
•    Generative AI should provide a similar quality of service for identified demographic groups, including marginalized groups. 
•    It should be used with careful consideration of the potential impacts, risks, and benefits of its use, and mitigation strategies to address them.
2. Transparency and accountability
•    Generative AI should be used in a transparent and accountable manner, with clear roles and responsibilities, oversight mechanisms, and documentation. 
3. Inclusion
•    Generative AI should be used in a respectful and inclusive manner, with due regard for the rights, dignity, and diversity of the people and communities affected by its use. 
4. Privacy and security
•    Generative AI should be used in a secure and trustworthy manner, with appropriate safeguards and controls to protect the data, systems, and outputs from unauthorized or malicious access, use, or modification. 
5. Innovation
•    Generative AI should help government more efficiently and effectively serve the public, with employees educated and empowered to consider uses for the technology that advances county’s mandate, objectives, and values.
- Policy requirements
a. Fairness and responsible use
1. Generative AI tool list
•    Department will develop and maintain a list of generative AI tools approved for government use. 
•    The list shall note the types of data that may be input into each tool based on the data classification policy. 
•    The department shall complete an impact assessment and security assessment prior to adding a tool to the approved generative AI tool list.
2. System development
•    When developing or deploying systems that include generative AI features, employees shall follow guidance developed by department to ensure generative AI tools or generative AI systems are fit for purpose, have appropriate human control, and provide consistent experiences across groups.
3. Review of AI output
•    Employees are responsible for the output generated by generative AI tools. 
•    Generative AI content shall be reviewed before it is disseminated. 
•    Generative AI features built into tools that interface directly with the public, output should be evaluated before the tool enters production and periodically thereafter
b. Transparency and accountability
1. Disclosure of AI interaction
•    Interactions with a generative AI tool must be clearly labeled and distinguished from human-generated content or data. 
•    This does not apply to content generated by AI that is reviewed or edited prior to sharing with others.
2. Impact assessment
•    Before deploying systems that include generative Ai features, government employees shall assess the impact of the system on people, organizations, and the public by completing an impact assessment using the template provided by the department, including identified risks and mitigation. 
3. Decision making
•    Government employees shall not rely solely on output from generative AI tools to make decisions that could have a significant adverse impact on individuals or communities, including impacts on civil rights, civil liberties or physical or mental health. 
4. Accuracy
•    Government employees shall verify and validate the quality and accuracy of the output of generative AI services before using it for decision making, communication, or dissemination.
5. Monitoring
•    Department shall subject generative AI tools and features to regular monitoring, evaluation, and audit to ensure its effectiveness, efficiency, and compliance with the policy principles and requirements.