Smartphones gave workplace misconduct cases something they never had before: receipts. Generative AI is taking them away, and proving what happened at work is now harder than it has been in decades.

For most of the history of these disputes, wrongdoing came down to word against word. The comment in the hallway, the hand on the shoulder, the threat behind a closed office door. No witnesses and no record, so outcomes turned on who sounded more believable. True accusations, no matter how credible, could die on the vine from lack of proof. Then work moved onto phones and chat platforms, and misconduct started creating records. The harassing message arrived with a timestamp and an author attached. The screenshot ended so many arguments that the culture gave it a name: receipts.

Both pillars holding up that era have cracked at once. A convincing fake thread takes minutes to make, and fabricated messages have already decided employment matters. I wrote recently about message screenshots arriving in court as evidence , including a federal employment case dismissed with prejudice after the court found the plaintiff's image of harassing texts was fabricated, and a Colorado disciplinary ruling that disbarred a former prosecutor over four fabricated texts framing a coworker for sexual harassment.

Meanwhile the authentic thread can now be denied with a straight face, the workplace edition of the liar’s dividend , where real recordings get waved off as AI. A receipt such as a text message exchange that can be forged in one direction and denied in the other stops settling anything on its own. Somebody has to check.

In a recent interview with Human Resources Director , Tracey Diamond, a labor and employment partner at Troutman Pepper Locke, called workplace deepfakes "a continuation of the old version of employees lying," with employees now "creating evidence to further their lies." She told the outlet she has already handled an investigation involving falsified communications. Her advice for a suspect message: "checking the employee's phone as well as the alleged harasser's phone to see if the messages are on both phones, checking the deleted files, performing forensics. Do the best they can to corroborate the story."

An employment lawyer is describing device-level digital forensics work as the standard for a category of dispute that, through most of its history, would not elevate to the level of forensic examination. This changes the dynamics of employee wrongdoing cases, and the costs.

The New BYOD Problem Has Nothing To Do With Cybersecurity

Receipts were never only text messages. They accumulate in email, in Slack and Teams channels, in chat apps and direct messages, anywhere work conversation happens. Where they accumulate decides how hard they are to verify. Inside the systems an employer runs, the records keep themselves: email archives, collaboration platforms and the audit logs behind them, maintained by machinery no single party controls. A conversation that supposedly happened there can be tested against records neither side can quietly rewrite. But these records do not account for personal devices.

For a decade, bring-your-own-device policy has been a cybersecurity conversation: protecting company data on phones the company does not own. The new BYOD challenge has nothing to do with cybersecurity. Those same phones carry the personal text threads, chat apps and direct messages where misconduct often happens. When a dispute erupts there, the employer's records have nothing to say, and the only road back runs through the original personal devices, reached through consent or through counsel.

Hardest On The People Telling The Truth

The new burden does not fall evenly. The employee with a genuine complaint now meets a doubt that barely existed a few years ago: maybe the thread is generated. The employee framed by a fake inherits the job of proving a negative. The coworker in the Colorado matter cleared his name because a digital forensic examination of his own phone , according to reporting on the ruling, found no communication between his number and hers. He was cleared by a device. Not everyone accused by a fake will have one to offer, or an employer willing to look.

Employers sit in the middle of the squeeze. Diamond's reminder in the same interview is that they owe a prompt and thorough investigation whenever a policy or a law may have been violated. The obligation has not changed. The evidence underneath it has. Act on a fabricated message and someone was disciplined for words they never wrote. Treat every message as possibly fake and do nothing, and real misconduct continues with a technical excuse attached.

Even if most workplace receipts are still real, the fabricated ones will look just as real. And this is where the peril is. The screenshot that ends a career, or protects one, now gets a treatment it may never have had before, a digital forensic examination of the smartphone or laptop that allegedly created it.

Deepfakes are creating this need. While it adds complexity and cost these cases rarely carried before, in another sense this is good news. Cases that never got device-level examinations now will, and device-level forensics in almost every instance is as close to proof as digital evidence gets.