AI, Slack, Teams, and the Modern Evidence Problem: Technology in Workplace Investigations
What Employers Need to Know
|
AI as the investigator's tool.
Used properly, AI addresses the modern investigation’s defining problem: volume. Workplace conduct now lives across email, Slack or Teams, texts, tickets, and shared documents, and the communications relevant to a single harassment complaint can run to tens of thousands of messages. Machine review—semantic search, pattern surfacing across channels, transcription and summarization of interviews—collapses weeks of review into days. Law firms, including ours, are building these capabilities directly into investigative practice when appropriate and permitted. The discipline is in the boundaries. AI surfaces and summarizes but humans verify and reach conclusions. Every machine-identified lead must be verified against the source by a person whose name goes on the report, because these tools miss context, misread sarcasm, and occasionally assert things that are inaccurate. Investigators should also use enterprise tools with confidentiality commitments and access controls, not free consumer chatbots. If your organization wants to use AI to assist in its investigations (or on other employment decisions), you also must remain mindful and aware that certain states, like California, have placed limitations on how and when employers may use AI in employment-related decision-making.
AI as the misconduct and the counterfeit evidence problem.
Unfortunately, the same tools that can help move an investigation along sometimes generate the conduct being investigated. Deepfake images and audio of coworkers have arrived in ordinary workplaces, and complaints about AI-generated harassment now require investigators to answer questions that did not exist five years ago. You now have to ask who created it, what tool did they use, how was it distributed, and is the artifact itself even what it appears to be? Investigators now have to conduct evidence authentication. The screenshot of a text conversation was once functionally self-proving. Today a convincing fabrication is a free app away and so is a false claim that authentic evidence was fabricated. Investigators need both instincts. They must corroborate the source of digital evidence (platform records and metadata over screenshots where the stakes warrant) and resist the reflexive fraud defense (“that's a deepfake”) from becoming an automatic assumption. Meanwhile, the quieter technology problems have compounded. There are ephemeral messaging apps whose entire design defeats preservation, BYOD phones where the company’s right to look is only as good as the policy nobody updated, and workplace monitoring tools with outputs that are admissible evidence in some states only if the required notices were given.
What can employers do now to evaluate their own AI use in investigations and to gain awareness around authentication?
- Adopt an AI-use policy for investigations. Outline the approved enterprise tools, prohibited consumer tools, and human-verification requirements.
- Update BYOD and acceptable-use policies to secure meaningful investigative access to business communications on personal devices.
- Add authentication practice to investigator training. Investigators need to know when to demand platform records over screenshots, and how to respond when a party claims evidence is synthetic.
- Extend your policy definitions of harassment to expressly cover AI-generated content.
Next, and final, installment: when the investigation becomes Exhibit A. How plaintiffs attack investigations in litigation and a self-audit checklist drawn from the full year.
The Investigation Roadmap: A Twelve-Part Series for Employers
This is an installment of the Investigation Roadmap, a twelve-part Vorys at Work series on workplace investigations. It is published weekly and walks through the full life of a workplace investigation, from the moment a complaint lands to the day the file becomes an exhibit.
