The Department of Labor (DOL) in the United States recently unveiled a comprehensive AI & Inclusive Hiring Framework aimed at guiding employers on how to integrate artificial intelligence into their hiring processes while avoiding discriminatory practices. This framework highlights ten focal areas designed to promote compliant and inclusive use of AI, providing detailed recommendations that employers could adopt to align with legal requirements and emerging standards in AI technology within employment contexts.
The framework’s first key area advises employers to identify and understand the legal requirements that pertain to their use of AI in hiring. The relevance of specific laws will vary based on factors such as the business's location, its number of employees, and any applicable federal or regional regulations. Given the rapidly evolving landscape of AI-related legislation, employers are also cautioned to remain vigilant about new legal requirements that might emerge.
Establishing specified roles and training within an organization is another focal area. The DOL underscores the importance of forming diverse teams to facilitate AI technology implementation, which can contribute significantly to creating an inclusive work environment. Organizations are encouraged to consider collaboration between various internal departments and leadership to effectively govern AI activities.
The framework further advises businesses to maintain an inventory of their AI hiring technologies. This involves gathering complete information on the AI tools in use and establishing clear policies concerning their acceptable applications — all contextualised within the perceived risks and operational capabilities.
Working with external vendors and third parties also forms a crucial part of the framework. Employers are advised to draft comprehensive policies for engaging with these vendors to ensure compliance during procurement and implementation phases, identifying risks associated with third-party software, and setting up clear reporting processes.
For a robust assessment of AI’s impacts, the framework recommends that employers promote a workplace culture that values risk-awareness. This involves conducting collaborative assessments involving internal staff, experts, independent evaluators, applicants, and current employees to gauge the AI's influence on inclusivity and compliance.
Additionally, the framework highlights the need for employers to facilitate a process for applicants to request accommodations if they feel uneasy with the use of AI in hiring. It should be noted, however, that no federal or state law currently requires employers to provide such accommodations automatically, though existing disability discrimination laws may necessitate it in particular situations.
The emphasis on using "explainable AI" calls for employers to seek clarity from vendors regarding how their AI systems function and to provide accessible notices outlining the technology's operations and purposes. This transparency is crucial in building trust and understanding among users and applicants.
Ensuring human oversight is another critical recommendation. Employers are encouraged to assign individuals who focus on diversity, equity, inclusion, and accessibility to key roles in overseeing AI technology, ensuring its proper usage and addressing any concerns that may arise.
The potential for incidents and failures in AI systems is acknowledged within the framework, which advises employers to establish responsive policies to handle such events. Procedures should also accommodate appeals from external users affected by AI shortcomings.
Regular monitoring of AI technology performance is vital, according to the framework. Systematic assessments are suggested to ensure adherence to trustworthiness, compliance, and non-discrimination standards, allowing employers to maintain an inclusive hiring process.
While the DOL’s AI & Inclusive Hiring Framework does not have binding legal power, it serves as a comprehensive guide for companies wanting to shield themselves against claims of AI-driven discrimination. Employers are encouraged to evaluate which recommendations best fit their business model and could be implemented with ease, without the need to adopt all suggestions simultaneously.
Source: Noah Wire Services