For many years, corporations have harnessed the facility of synthetic intelligence (AI) in numerous enterprise functions. However, the newer utilization of AI instruments in employment selections has intensified, igniting issues about bias and discrimination towards job candidates and workers.
Since selections made by AI are solely nearly as good because the algorithms and information they depend on, decision-making could also be influenced by biases inherent within the algorithms and information used to generate the outcomes. Sometimes this leads to disparate or hostile impacts on protected teams.
The Equal Employment Opportunity Commission forbids employment practices which have a disparate affect on protected courses of employees. It just lately issued a technical help doc specifying that the usage of AI instruments for employment selections is taken into account a variety process and that an hostile affect on a protected class brought about using an AI software is a violation of Title VII of the Civil Rights Act, which prohibits employment discrimination.
The steering additionally indicated that an employer will probably be held accountable for employment actions taken based mostly on selections made by an AI software even when the AI software was developed by an outdoor entity.
This doesn’t essentially imply that employers ought to cease using AI instruments altogether. When AI instruments are used fastidiously, they will really cut back bias and discrimination in employment decision-making. Employers that need to act responsibly and avoid litigation ought to implement the next steps.
Human oversight and accountability
A particular particular person or staff ought to be assigned the accountability of monitoring and approving the usage of AI. These workers ought to be educated concerning the potential for bias, how to counteract it, and the implications of hostile impacts on protected courses. The finest apply is to require that each one selections based mostly on AI suggestions be made by a human.
Establish insurance policies for AI use
Policies ought to embrace phrases that restrict utilization for employment selections, require workers to disclose when they use AI to full a job, make clear expectations for information privateness, confidentiality, and safety, and element penalties for workers that violate the insurance policies.
Training and schooling on AI
Educating workers on the approved use of AI and the potential for biased outcomes can be utilized to fight irresponsible use of AI. Employees shouldn’t be allowed to use AI for enterprise functions with out understanding the potential for disparate or hostile impacts towards protected teams. Training ought to embrace ideas for decreasing or eliminating bias within the outcomes.
Implement moral AI instruments
(*5*) ought to prioritize the implementation of moral AI instruments. Companies that provide moral AI instruments are trustworthy concerning the potential for bias inherent within the information. They are additionally clear about how they establish biases and cut back their affect on decision-making. Given the sensitivity of personnel information, companies ought to search AI corporations with strong practices relating to information privateness and safety. If it isn’t instantly apparent on their web sites, corporations ought to test the phrases and situations earlier than using them.
Related: 65% of HR managers say using AI instruments will probably be a high ranked talent, research finds
Regular audits of employment selections
Finally, corporations ought to conduct common audits of their employment practices together with the usage of AI to be certain that they aren’t making a disparate or hostile affect on a particular protected class.
Stefanie Camfield is an assistant common counsel and human sources advisor at Engage PEO.
https://www.benefitspro.com/2023/09/05/5-ways-for-employers-to-avoid-discrimination-when-using-ai/
