Key Points
- As employers and employees are increasingly using AI in the workplace, government regulation is playing catch-up.
- Existing EEO laws place some guardrails on an employer’s use of AI for traditional human resources tasks, and Illinois and New York City have enacted laws specifically regulating the use of AI in employment decisions.
- Employers should monitor for updates and mitigate risks when using AI systems. That may entail training HR and other staff and, in some cases, commissioning audits of the impact of the systems.
The use of artificial intelligence (AI) in the workplace is growing exponentially. Companies are looking to harness AI to improve productivity and efficiency as AI’s capabilities continue to expand. In doing so, employers should be mindful of the evolving legal landscape surrounding the use of AI, including existing employment-related laws.
Artificial Intelligence
AI generally refers to a computer performing tasks historically done by a human. Generative AI refers to AI algorithms that create new content based on existing data. AI generally draws on existing data and practices that are used to train the technology. (See our Spring 2023 The Informed Board article “What Is Generative AI and How Does It Work?”)
In the workplace, AI can be used for many functions, including recruiting and hiring. It can streamline these processes by, for example, preparing job descriptions and analyzing internal employer data to best predict which applicants would be most successful at the company. AI systems can be used to analyze productivity, measure individual performance and select candidates for promotion.
AI proponents argue that AI reduces or eliminates human bias in screening and evaluating applicants and employees. However, because AI technologies generally draw on existing data and practices, some argue that use of AI for these functions may perpetuate any existing discriminatory practices.
In addition to human resources functions, employees are also using chatbots to augment their work. For example, employees can outsource time-consuming tasks such as drafting an email, writing code or doing research to an AI chatbot. AI proponents believe it will revolutionize the workforce and reduce — if not eliminate — the need for humans to perform certain tasks. But information provided by chatbots may be incorrect or outdated because chatbots survey and pull from troves of publicly available information, which may themselves be incorrect or outdated.
Chatbots also “hallucinate” or make things up. There are also privacy concerns associated with use of AI, as information (which may include confidential or proprietary information) entered into a company’s AI system might be inadvertently disclosed in the output given to a third party or, if such information is submitted in a query to a chatbot, it might be incorporated into the chatbot’s database and revealed in a subsequent response to an unrelated party’s query.
Existing Guardrails
An employer’s use of AI must comply with existing employment laws, including the following that prohibit discrimination:
- Title VII of the Civil Rights Act of 1964 (Title VII) — on the basis of race, color, religion, sex or national origin.
- Americans With Disabilities Act (ADA) — against individuals with disabilities.
- Age Discrimination in Employment Act — against applicants and employees who are 40 years of age or older.
- Similar state and local laws.
Both federal and state laws prohibit disparate treatment discrimination in the workplace (i.e., intentional discrimination of members of a protected class) and disparate impact discrimination (i.e., facially neutral practices or policies that may disproportionately affect members of a protected class).
To the extent an employer uses AI in connection with recruitment, hiring and promotion, there is a risk that the employer may run afoul of applicable anti-discrimination laws. For example, if the AI model bases hiring decisions on criteria such as ratings, pay and titles that skew in favor of white or male applicants, the AI system may screen for applicants meeting those characteristics.
There is also a risk that AI may make inferences based on an applicant’s religion, age, sexuality, genetic information or disability status learned from either the internet, social media or the applicant’s resume or interview. Though AI may not intentionally screen out or favor certain protected classes, AI’s choice could cause a discriminatory impact.
In the event that use of AI results in employee layoffs, employers should also ensure that they are mindful of anti-discrimination laws when selecting those who will be separated from employment. They must also pay close attention to the requirements of the federal Worker Adjustment and Retraining Notification and equivalent state and local laws, which govern notice obligations in connection with plant closings and mass layoffs.
Additionally, to the extent employers use AI in robotic systems or other machinery, they should be sure to comply with the Occupational Safety and Health Act and equivalent state laws and regulations, which require that employers provide a workplace free from recognized hazards.
New AI Legislation and Guidance
Increasingly, federal, state and local governments have considered legislation to supplement existing employment laws and taken other steps in response to the rapid increase in use of AI in connection with employment.
Federal Laws and Action
Although no comprehensive federal legislation has passed, there are several bills in Congress addressing different aspects of AI. For example, the Algorithmic Accountability Act of 2022 bill directs the Federal Trade Commission (FTC) to require entities to conduct impact assessments for bias, effectiveness and other factors when using automated decision systems to make critical decisions.
Additionally, the White House has released a Blueprint for an AI Bill of Rights and an AI Risk Management Framework, and President Joe Biden signed an executive order related to bias and discrimination with respect to the use of AI. (For more on the AI Risk Management Framework, see our May 18, 2023, client alert.)
On May 12, 2022, the Equal Employment Opportunity Commission (EEOC) issued technical guidance addressing the use of AI to assess job applicants and employees under the ADA. The guidance outlines several ways that utilizing AI tools can violate the ADA, including, for example, using AI to evaluate video interviews without providing notice and the opportunity for applicants to request a reasonable accommodation, in light of the potential for AI to negatively evaluate a candidate with a speech disorder.
On May 18, 2023, the EEOC issued technical guidance addressing the use of AI to assess job applicants and employees under Title VII. The guidance outlines how an employer’s use of AI tools can violate Title VII under a disparate impact analysis. It recommends that employers consider, among other things, whether the software has been evaluated to ensure it does not result in substantially lower selection rates for individuals with protected characteristics.
State and Local Laws
At the state level, Illinois enacted the Artificial Intelligence Video Interview Act in 2022, which imposes certain requirements on employers that analyze video interviews with AI. Many other states, including California, New Jersey, New York, Vermont and Washington, D.C., have proposed or are otherwise considering legislation to regulate AI use in hiring and promotion.
At the local level, New York City Local Law 144, which sets forth limitations and requirements for employers using automated employment decision tools (AEDTs) to screen candidates for hire or promotion, will be enforceable as of July 5, 2023.
- The law prohibits use of AEDTs unless the tool has been the subject of an independent bias audit within the past year. Employers and employment agencies are also required to make public on their websites the date and a summary of the results of the most recent bias audit and the AEDT’s distribution date.
- In addition, employers and employment agencies must provide a notice of the use of AEDTs to employees and candidates for employment who reside in New York City. The notice must include instructions for requesting an alternative selection process or reasonable accommodation under other laws.
Employers in violation will be liable for a civil penalty of up to $500 for a first-time violation and between $500 and $1,500 for successive violations.
Best Practices
Against this backdrop, employers should consider implementing policies and procedures governing employees’ use of AI at work. When using AI technologies for hiring, screening and promotion, employers should consider auditing the technology to ensure selection rates do not violate anti-discrimination laws, to the extent employers are not otherwise required by applicable law to conduct audits.
Employers should also consider clearly informing applicants about the type of technology that will be used and the information it is measuring. Training staff regarding their use of — and potential discriminatory issues associated with — AI tools and the need to verify the accuracy of any AI-generated content is also advisable.
See all of Skadden’s June 2023 Insights
This memorandum is provided by Skadden, Arps, Slate, Meagher & Flom LLP and its affiliates for educational and informational purposes only and is not intended and should not be construed as legal advice. This memorandum is considered advertising under applicable state laws.