As deepfake technology becomes more accessible, employers are facing new risks associated with the use of artificial intelligence and workplace conduct. In this episode of “The AI Didn’t Know It,” Employment Vice Chair Julie Levinson Werner talks to partner Brian Sterba about how AI-generated content can give rise to harassment, discrimination, and hostile work environment claims. This discussion highlights the evolving legal landscape, including recent federal and state developments, and outlines practical steps organizations can take to update policies, train employees, and respond effectively when issues arise.
Partner Brian Sterba explores the myriad legal and business implications of this innovative technology. Brian and guests from a variety of corporate practices will discuss the different ways companies are leveraging AI to improve productivity and efficiency, as well as the challenges that arise as companies, Congress, and regulators weigh different approaches to managing intellectual property rights, privacy and data security, national security interests, and more.
Brian Sterba: Welcome to another episode of “The AI Didn’t Know It.” Today my partner, Julie Werner, and I are discussing the intersection of AI and workplace harassment.
Julie, we’re hearing more and more about artificial intelligence in the workplace, and one area that’s gaining attention is the use of deepfakes and other realistically manipulated content. Employers and HR teams need to really understand this issue because it’s not just about technology, but also about employee relations, compliance, and legal risk.
So why is this issue important now, and what can or should employers do about it?
Julie Levinson Werner: Thank you, Brian.
This is a very important topic because deepfakes are no longer just a theoretical concern or a problem that occurs somewhere outside the workplace. These tools can create highly realistic images, videos, and audio, but if that content is targeted at employees or distributed in the workplace, it can lead to very serious legal and employee relations issues.
Brian Sterba: Let’s start with the big picture. When people hear the term “deepfakes” they may think of it as a technical issue or just an online issue, but why is this becoming a real workplace issue for employers and HR professionals?
Julie Levinson Werner: This is a workplace issue because the harm goes beyond the technology itself. When deepfake content involving employees is created or shared in the workplace, it can impact that employee’s dignity, well-being, and ability to work.
The EEOC has already warned that sharing deepfakes and other AI-generated images and videos in the workplace may constitute harassment based on legally protected characteristics. Therefore, this can quickly become an issue of not only misuse of the technology, but also harassment, discrimination, and a hostile work environment. And from an HR perspective, that means employers need to be prepared to respond in the same way they would respond to other serious workplace misconduct.
Brian Sterba: What are the main risks for employers? Are we primarily talking about harassment and hostile work environment concerns, or do the risks go beyond that?
Julie Levinson Werner: Claims of harassment and hostile work environments are certainly central, but the risks can go beyond that. We have already seen examples of employees suing their employers based on allegations that deepfake content was created or distributed by co-workers and the employer failed to respond appropriately.
What stands out in these cases is that the claims often focus not just on the existence of deepfakes, but also on what the employer did or did not do after learning of the issue. Employees have made allegations of sexual harassment, sexual discrimination, and a hostile work environment, which are common allegations, as well as claims of negligent retention, negligent supervision, retaliation, and failure to prevent the spread of content.
Therefore, one of the biggest risks for employers is not just that something will happen, but that the organization fails to act quickly, seriously and effectively after becoming aware of a problem.
Brian Sterba: How is the legal landscape evolving on this issue, and what should employers be paying attention to now when considering both federal and state trends?
Julie Levinson Werner: At the federal level, an important new framework, the Take It Down Act, will be implemented, making it a crime to knowingly publish certain AI-generated intimate visual depictions, including digital forgeries, without consent.
A federal law called the DEFIANCE Act has also been proposed. The DEFIANCE Act creates a civil remedy system for victims, allowing them to seek substantial damages and attorney fees, and in some cases file lawsuits under false names.
There are also laws targeting deepfakes at the state level, adding further complexity for employers operating in multiple jurisdictions.
In addition, there remains uncertainty as developments in federal policy may impact how some states handle AI laws going forward. Therefore, if your employer is responding to an area where the law is rapidly evolving, this issue may remain unresolved for some time.
Brian Sterba: So, from a practical perspective, what can employers do if they want to be proactive rather than reactive?
Julie Levinson Werner: The first step is to make sure your workplace policies are up to date and clear. Employers can have policies that explicitly address the creation and distribution of deepfake content and make clear that this type of behavior may violate harassment and misconduct policies. It may also be grounds for disciplinary action, including dismissal.
Training is also important. HR professionals need to know how to respond to these complaints, and managers need to understand that when they learn of an issue, they need to immediately escalate the issue rather than handling it informally or dismissing it as a joke.
It is also important to have clear reporting and response protocols in place. Employers should think ahead about who needs to be involved when a complaint comes in, such as human resources, legal, or even IT, so that the organization can respond quickly and consistently.
Brian Sterba: And if an employee reports manipulated content circulating about them, what should an employer do first? What should they do now?
Julie Levinson Werner: Initial responses must be swift, serious, and cooperative. Employees need to know their complaints are being taken seriously, and organizations need to act quickly to preserve evidence and understand what happened. This typically means documenting the complaint, preserving relevant materials, identifying where the content was shared, and coordinating between HR, legal, and IT as necessary.
If another employee may be at fault, employers should promptly investigate, gather relevant evidence, interview parties and witnesses, and take corrective action as appropriate.
Employers should also consider employee support and practical containment. So, for example, while a human resources department may not be able to remove content from the internet on its own, an employer can notify employees of their right to request removal under applicable law. You can also work internally with your IT department to restrict distribution on corporate systems and networks.
What employers should not do is minimize complaints or delay action. In this area, inappropriate responses can represent a significant portion of the legal risk.
Brian Sterba: As employers think more broadly about AI, what are the key points you want HR professionals and business leaders to keep in mind regarding deepfakes in the workplace?
Julie Levinson Werner: Employers need to treat this as a current workplace risk, not a hypothetical for the future. Deepfakes can cause serious harm to employees and pose significant risks to organizations, especially if employers are unprepared to respond. The best approach is for companies to be proactive. Update policies, train the right people, establish reporting and response procedures, and stay informed of legal developments.
After all, the most important thing here is preparation, speed and judgment.
Brian Sterba: Julie, thank you for that thoughtful overview. We hope to help employers understand both the legal risks and practical steps their organizations should consider in this area.
Next time, please join us for “AI Didn’t Know That.”
