BREAKING: The Culture Shifters Annual Report is here!
Get your copy.

AI Is Changing Workplace Harm Faster Than Employers Can Respond

2 min read
Published on
January 22, 2026
From manipulated images to synthetic video, deepfakes are no longer a distant or theoretical risk. They are already beginning to show up in workplace contexts, creating new forms of harassment that are harder to detect, harder to prove, and significantly more damaging for those targeted.

Our CEO and Co-founder, Gemma McCall, recently explored this issue in Personnel Today, following growing concern around the misuse of AI-generated content in professional environments. Here is a taster of what Gemma discussed:

“Deepfakes raise serious legal and compliance concerns for employers. HR should treat the risk as high where deepfakes are used to humiliate, control or punish.”

For employers, this shifts the conversation. This is no longer about whether something happened in a physical workplace or online. The distinction is disappearing, and so is any protection that distinction once offered.

Legislation is already catching up. The Online Safety Act has expanded how online harm is defined, and expectations around employer responsibility are evolving alongside it. If an organisation fails to respond appropriately to AI-enabled harassment, the consequences are not abstract. They include reputational damage, loss of trust, and the very real possibility of tribunal claims.

“How HR teams respond to the emerging threat of AI-enabled harassment will be a test of culture, not just compliance.”

This is where many organisations will be exposed. Policies written for a pre-AI world will not stand up to incidents involving synthetic content. Investigations will be more complex. Evidence will be harder to interpret. And the impact on individuals can be severe, particularly where content is designed to humiliate or undermine.

Under the Worker Protection Act, the steps an organisation takes to prevent harassment are already under scrutiny. In the context of deepfakes, those steps will be judged not just on process, but on intent. Waiting until an incident occurs is no longer a defensible position.

Read our full article in Personnel Today to understand what this shift means for your organisation and how HR leaders should be preparing now.

Hannah Whitby
Head of Marketing
STORIES

What we're reading

Latest insights from the front lines of workplace culture.

Blog
4 min read

What Happens Before a Misconduct Panel? Lessons from the University of Sussex

At a recent Community of Practice session, colleagues from the University of Sussex shared how they've strengthened their student misconduct processes. One insight stood out: successful outcomes don't start in the hearing room. They start long before a panel is ever convened.

Read more
News
3 min read

The Tribunal Risk HR Leaders Can't Ignore

Featured in HR News, our latest research in the Culture Shifters Annual Report 26-27 reveals a stark gap between how confident HR professionals feel about sexual harassment compliance and what they can actually prove. Of 300 senior UK HR professionals surveyed, 91% said they're confident their organisation could evidence taking "all reasonable steps" to prevent sexual harassment if challenged today. But when asked to demonstrate specific measures, the evidence often isn't there. Only half can show clear reporting processes are in place. Only a third have measures to prevent retaliation.

Read more
Webinar
45 min watch

The Hidden Costs of Workplace Misconduct in Financial Services

Join Ash McDowell, Chief Financial Officer at Culture Shift, and Christine Bonney, Head of Culture Transformation, as they explore the true cost of workplace misconduct in financial services and what organisations can do to reduce both risk and cost.

Read more
CULTURE SHIFT

Feeling inspired?

Take the first step toward preventative misconduct management with a demo of our Report + Support™ platform. We can show you how to breakdown reporting barriers with anonymous 2-way messaging, and how to act before things escalate with name-matching and pattern-spotting across our analytics dashboard.

Dotted background image