Businesses have officially moved from the “experimentation” stage to the “integration” stage of utilizing artificial intelligence (“AI”) in business operations. AI tools are increasingly being used in the entire lifecycle of the employment relationship, from recruitment to termination and everything in between, including such things as performance management and scheduling. While these tools can maximize efficiency, they present many legal considerations that companies should be aware of.

It is important to understand that the use of AI in any capacity compromises privacy. As such, the use of AI tools in managing any stage of the employment relationship implicates privacy laws which govern collection, processing, and storage of personally identifiable information (“PII”). Unfortunately, the United States does not have a unified legal framework when it comes to privacy law but instead has a patchwork of state and federal laws that may or may not apply depending on the industry, methods of collection, reasons for processing, and means of storage. Depending on the type of PII, your business’s sector, and the nature of your use of the data, you could implicate federal laws like HIPAA, the Electronic Communications Privacy Act, and the Fair Credit Reporting Act, or state laws like the Florida Security of Communications Act or the California Consumer Privacy Act.

On top of implicating the laws more typically associated with privacy concerns, the use of AI tools in employee management implicates traditional employment laws. Risk for employers can arise due to algorithmic bias in the AI tools. Algorithmic bias occurs when machine learning results in algorithms that produce discriminatory outcomes. Algorithmic bias is often unintentional and based on unknown biases in the data that becomes part of the algorithm. For example, if a company’s historical recruitment data reflects that more men than women get past the resume vetting process, that skew will be baked into the algorithm regardless of whether the reasons for the imbalance are legitimate.

If businesses are making employment decisions based on the AI tools they are using, they will not be able to explain the reason for those decisions—reliance upon an algorithm does not give employers a pass on complying with anti-discrimination laws.

Additionally, some jurisdictions, such as New York State, outright prohibit the use of “automated decision-making” in employment. Although Florida does not currently have a law like this, depending on the jurisdiction, Florida employers could be liable in other states if they are using AI tools to make employment decisions that affect residents of other jurisdictions.

Any employers using AI tools in employment decisions should be mindful to:

  • Provide proper notice and get appropriate consent before anyone’s information is provided to the AI tool.
  • Conduct audits of AI outcomes as well as data inputs to evaluate for bias.
  • Negotiate your contracts with the companies providing the AI tools to make sure that you have favorable allocation of liability and a clear understanding of what is being done with your employees’ data.
  • Never fully delegate any employment decision to an AI tool.

Even ChatGPT agrees with us on this – when queried about the risks for employers using AI in employee management, it cautioned of legal, operational, and confidentiality risks and provided similar recommendations, a couple of which are copied here:

  • Employers should ensure that all AI-generated employment decisions are reviewed by a trained human decision maker. “Human in the loop” review is becoming increasingly important from both a compliance and litigation-defense perspective.
  • Employers should be mindful that AI-generated communications or evaluations may later become discoverable evidence in litigation and may be scrutinized for bias, inconsistency, or lack of individualized assessment.

Stearns Weaver Miller’s privacy team is fully equipped to guide your company through any questions or issues that you have regarding privacy, AI, or how they affect your employment or larger business practices.