CXR RFI/RFP AI Template & Guidelines

Definition:

1. What is your definition of AI with regards to the solution your product(s)/service(s) offers?
   
   

Quality Control [Internal/External]:

1. How do you ensure the quality and relevance of the data used to train your generative AI model for talent acquisition and hiring?
   
   
2. Do you contract with an independent technology service to verify your AI claims? If you do, please supply their reports that include a description of their credentials, criteria used, methodology and assessment.

Operational Considerations [Scalability, Security, Privacy, Compliance, Training, Documentation]:

1. Describe your methodology for collecting, processing, and verifying data from resumes, job descriptions, feedback, interviews, or other data sources relevant to your solution.
   
   
2. What are the sources of data on which your AI is trained and to what degree to you have the rights to this data?
3. Describe how your software handles large data sets and high volumes of requests.
   
   
4. Describe how your solution handles and protects sensitive personal data aimed at protecting end users’ privacy. Provide specific methods used to anonymize data, how data retention policies are administered including the right to be forgotten, and any other measures taken to protect users’ privacy.
   
   
5. Describe your approach to complying with relevant laws and regulations in multiple jurisdictions, such as GDPR, CCPA, or EEOC?
   
   
6. Are your algorithms, as used in your solution, static? If not, do they continue to learn within our use or, more broadly? How do you support your solution over time?
   
   
7. What Information Security policies and procedures do you have in place? At a minimum, please share your policies for data protection, encryption, access control, audit logging, and incident response.
   
   
8. How do you support the integration and customization of the generative AI solution for talent acquisition and hiring?
   
   
9. Can you adjust the creativity, level of detail, and style of the generated outputs?
   
   
10. Is it transparent and explainable?
    
    
11. What level of ongoing support and maintenance is included? How are updates and bug fixes handled?
    
    
12. How do you provide documentation, training, and technical support for users?

Bias Mitigation/Misuse:

1. How do you address the issues of data diversity and bias in talent acquisition and hiring?
   
   
2. How is human input used to guide & refine the generative process, integrated into recruiter workflows? Is there an inbuilt “verify & override” workflow built into the design?
   
   
3. How do you prevent or mitigate potential harm or abuse of generative AI, such as plagiarism, inaccurate or misleading information (i.e. hallucinations) or manipulation in talent acquisition and hiring? Describe the guardrails you have put in place.
   
   
4. How will you support us if we are audited?
   
   
5. How do you address the ethical and social implications of using generative AI for talent acquisition and hiring?

Outcomes:

1. How do you measure and evaluate the effectiveness, accuracy, and reliability of the generative AI model for talent acquisition and hiring?
   
   
2. What are the metrics and benchmarks used to assess the quality and suitability of the generated content, such as candidates, profiles, messages, and reports?

Unintended Consequences:

1. How do you handle errors, anomalies, or inconsistencies in the generated content or the hiring process?
   
   
2. Are there any judgements pending against current, previous clients or yourself with respect to your product/service?