Late last year, President Biden issued Executive Order No. 14110 laying out his Administration’s policy directions to federal agencies concerning the use of Artificial Intelligence (“AI”) by individuals and organizations under their jurisdiction.
In response to the Executive Order, the Department of Labor (“DOL”) published a list of “Principles for Developers and Employers” (“DOL AI Principles”), establishing eight (8) principles to guide the development and integration of AI into labor relations between employers and employees.
The DOL’s AI Principles recognize the costs and benefits associated with the use of AI. While employers across a wide range of industries may be able to incorporate powerful AI tools and systems into their work to reduce costs and improve efficiency, the use of AI without clearly articulated principles and safeguards may have negative consequences for employees and, if unlawful, for the employer.
The Executive Order and DOL’s AI Principles are instructive for employers as they provide insight as to how the DOL may regulate this emerging technology in the future as well as how other federal and state agencies may approach its use by employers in other contexts.
DOL’s AI Principles
The DOL’s list of principles for AI developers are instructive for employers that are considering the use of these tools in the workplace:
1. Centering Worker Empowerment: Workers and their representatives, especially those from underserved communities, should be informed of and have genuine input in the design, development, testing, training, use, and oversight of AI systems for use in the workplace.
The DOL expressly identified this principle as its “North Star,” conveying its importance relative to the other principles.
This principle encourages employers using AI tools and systems to solicit participation from those who may be affected by their use of AI. The DOL advises employers to involve employees and the employee organizations that represent them in decision-making regarding the use of AI in the workplace, mirroring the Executive Order’s directive that “all workers need a seat at the table, including through collective bargaining.”
For agencies with represented workforces, this principle implicates the public employer’s existing legal obligation to meet and confer on mandatory subjects of negotiation and to provide notice and an opportunity to bargain the negotiable effects of a management decision.
As the “North Star” of the DOL guidance, employers should keep this principle front of mind when considering making a decision concerning the use of AI in the workplace.
2. Ethically Developing AI: AI systems should be designed, developed, and trained in a way that protects workers.
The DOL’s second principle builds on the first and emphasizes protecting workers from the negative consequences of AI.
Employers may play a part in the design, development, and training of AI for use in their workplaces. Accordingly, employers should closely review what data will be provided to the AI system and how the system will use and learn from such data. Importantly, employers should be mindful not to use data that promotes and reinforces biases and to safeguard confidential or privileged information about their employees.
Further, employers should be wary of arrangements that involve the sale of the employer’s data as such sales may be unethical and violate employees’ privacy rights.
3. Establishing AI Governance and Human Oversight: Organizations should have clear governance systems, procedures, human oversight, and evaluation processes for AI systems for use in the workplace.
The DOL’s third principle makes clear the importance of human involvement and oversight in the deployment of AI in the workplace.
Given that the advancements in AI are currently outpacing efforts to regulate the technology, the decision to implement AI in the workplace should be preceded by the implementation of safeguards to monitor its initial and subsequent use, so that the system acts in compliance with the employer’s intentions and any legal obligations related to its use.
4. Ensuring Transparency in AI Use: Employers should be transparent with workers and job seekers about the AI systems that are being used in the workplace.
The DOL’s fourth principle encourages transparency from employers regarding their use of AI tools and systems.
This principle mirrors emerging legislative efforts to develop a system that would require employers to provide notice when using AI to make an employment decision or to use the individual’s data or information as part of an AI system. For example, Assembly Bill (“AB”) 2930, which is pending in the California legislature, would require employers to notify applicants for employment if the employer intends to use AI as part of the employer’s hiring decision. The bill would also require that employers accommodate requests by applicants to opt out of the employer’s use of AI in such decision-making.
While the requirements under pending legislation are not yet law, it would be good practice for employers to develop policies and practices that increase transparency concerning their use of AI to make employment decisions.
5. Protecting Labor and Employment Rights: AI systems should not violate or undermine workers’ right to organize, health and safety rights, wage and hour rights, and anti-discrimination and anti-retaliation protections.
The DOL’s fifth principle is intended to ensure that employers use AI tools and systems in a manner that complies with existing legal obligations, whether labor or employment law.
One of the principal concerns related to the use of AI in employment decision-making is algorithmic discrimination; where an AI system uses biased data and information in such a way that perpetuates discrimination. Since AI systems use data and information from the real world to make decisions, biases present in the real world may, if unaddressed, result in outcomes that are themselves discriminatory.
To mitigate the risks associated with algorithmic discrimination, the Equal Employment Opportunity Commission (“EEOC”) issued guidance encouraging employers to audit their AI tools and systems in order to ensure that such tools and systems do not result in disparate (i.e., discriminatory) impact, which is prohibited under existing state and federal law.
Employers should be mindful about the adoption and use of new AI tools and systems and how those tools and systems are operating, as well as existing tools and systems that are beginning to incorporate AI into their products.
6. Using AI to Enable Workers: AI systems should assist, complement, and enable workers, and improve job quality.
The sixth AI principle is to enable employees to use AI to improve their work. This principle aligns with the overall goal of the Biden Administration and the DOL as they relate to the use of AI in employment contexts. Instead of using AI to reduce the number of workers and to replace employees with AI technology, this principle suggests that employers should look for opportunities where AI can assist employees in the performance of their duties.
7. Supporting Workers Impacted by AI: Employers should support or upskill workers during job transitions related to AI.
Similar to the other principles, the seventh AI Principle recommends that employers train employees on new AI systems, so that employees can learn and develop advanced skills that will make them more capable, competent, and competitive employees. Instead of replacing an employee with AI, the DOL encourages employers to support employees in gaining the experience and understanding, skills, and knowledge required in order to maintain and operate new AI systems.
8. Ensuring Responsible Use of Worker Data: Workers’ data collected, used, or created by AI systems should be limited in scope and location, used only to support legitimate business aims, and protected and handled responsibly.
DOL’s eighth and final principle relates to employee information and data.
As discussed above, employers should preserve and protect employee information and data and not compromise employee privacy or confidences through or by the use or sale of such information.
This principle also makes clear that employers should only use employee information for legitimate business purposes and not use it for other private purposes.
What’s Next?
Employers should stay tuned for guidance issued by other federal and state governmental agencies concerning the use of AI by employers in employment contexts. Relatedly, there are several bills currently pending in the California Legislature that may affect how employers may use AI.
Liebert Cassidy Whitmore is monitoring statutory and regulatory authority that may affect how employers may lawfully use AI as well as informal guidance, such as that issued by the DOL, and will be providing further updates as necessary.