Employers, of course, must understand how AI can be used for human resources functions such as hiring, monitoring and performance assessments; to help maintain the competitive edge of the organization; and the legal risks associated with the use of AI. Employees must be aware of the standards of the organization for the use of AI and how to avoid creating data security or confidentiality risks, while understanding how to use AI in accordance with employer’s policies and practices.
Regulatory context
As we have pointed out here And hereSpecific AI regulations is relatively limited, even in states such as California, Colorado and Virginia, which tend to be at the forefront of the regulation of new technologies. Most of the new legislation aims to modify or supplement existing laws. Although the majority of legislation is aimed at “high -risk systems” and the prevention of “algorithmic discrimination”, the extent and complexity of the problems have exceptionally made political decision -makers to achieve consensus on how to regulate AI.
At the federal level, Congress has not shown no indication that it will pursue a radical AI regulation. However, barely two weeks ago, nestled in the American house representatives, the increases in the Committee for the Budget Reconciliation of Energy and Trade were a provision which would prohibit the States from enforcing any AI law or regulation for ten years after the promulgation. This moratorium adopted the Chamber on May 22, but with slight changes intended to ensure that the provisions were focused on interstate trade. It is not clear that the moratorium will go to the Senate, but if it is adopted, it is almost certain to face challenges immediately on interstate trade and the reasons for the tenth amendment. Thus, it seems that Congress is willing to influence federal policy on AI in one way or another, and employers should continue to monitor developments.
In the context of employment, a large part of the regulation that exists concerns the use of AI in recruitment and hiring processes. One of the first laws regulating the use of AI in user contexts was New York Local Law 144And since its promulgation in 2023, a certain number of states have imposed limits on the use of AI for hiring – notably Colorado, Illinois and Maryland. State laws generally focus on transparency and avoidance of discrimination. Companies covered by laws are generally required to provide certain disclosure on their use of AI and obtain prior consent of employees or candidates concerning the use of AI in hiring.
AI challenges are unique to employers
The boom in remote work has led to new challenges mainly linked to the monitoring of the workforce. Employers who use activity trackers must ensure that they do not violate any confidentiality regulations for AI or data. AI and data confidentiality problems also arise when scannings and physical badge technology monitor employee movements are used to prepare performance assessments and make job decisions.
In addition to the confidentiality of employees, employers must be concerned about their own privacy – in other words, the confidentiality of business information or company staff, including employee medical information. Employers should train employees on the importance of avoiding inappropriate disclosure when using AI.
Employers must consider the possibility that the use of AI to filter candidates can lead to complaints of discrimination. Algorithms can be based on older data that can perpetuate discriminatory job decisions. For example, an algorithm can be based on data from an era when employees of a certain employment category were mainly white or male. In the affirmative, the algorithm can select candidates who are white men on the basis of these historical data, which can lead to allegations of discrimination by the female or minority candidates who are detected before the employer is even aware of their existence. Screening of AI applicants has also been subjected to the elimination of older candidates and disabled candidates.
With current employees, AI monitoring tools may not take reasonable accommodation into account. For example, an employer can use an AI tool to monitor and follow employee productivity, and the tool signals atypical breaks but is not programmed to consider housing. Consequently, the tool can report an employee for disciplinary measures to take excessive breaks, even if the employee has received additional break time as a reasonable accommodation for a handicap. As we have already reportedCalifornia Privacy Protection Agency has published regulations offered to solve this problem. The comments are always accepted on the proposed regulations.
The above illustrates that employers must ensure that they are not too linked to AI tools. A human still has to ensure that AI tools operate with precision and validate the results. Otherwise, the employer could be responsible under the fight against discrimination or other employment laws. If the employer uses AI tools provided by a supplier, he should understand that both The employer and the seller can be responsible for any violation. Thus, the employer must exercise appropriate monitoring and also consider including compensation provisions in his contract with the seller.
Suggestions for employers using AI
While employers continue to adopt and integrate AI, there are three fundamental elements that they should set up and maintain:
- Name an interfunctional group with the power to supervise all AI problems in the organization – in other words, to manage the AI in depth for the whole of the company.
- Define the basic rules for the use of AI, for example, by adopting an “acceptable use of AI” policy on a business scale, or by developing a complete governance framework to manage and assess the compliance of AI, perhaps on the basis of the NIST AI Risk Management FrameworkAnd offer continuous training to employees.
- Create a risk assessment process that evaluates AI uses both before and after deployment, which should be periodically updated to integrate any modification of the uses of the AI employer, risk tolerance or regulatory developments.
