By Phil Soar, co-founder of CloserStill Group and Chairman of Nineteen Group
The Times recently reported how a fake avatar beat 261 other applicants to get a job with a tech firm. And I should make clear that I am using The Times story as the basis for this article and credit Megan Harwood-Baynes for bringing it to everyone’s attention.
There is an exhibition connection as the story was told at RX’s Infosecurity and backs up superbly the developing obsession about how face-to-face, and trade shows, have a tail-wind in the world of AI.
The expert, who used widely available AI tools to pose as a woman of Asian heritage, is Jake Moore, who spent 14 years policing cybercrime. He used common tools to transform his face and voice into a young Asian woman (he did the same with a white man roughly his own age, 44, who went through another interview process and also got a job offer).
When deceiving, use a generic name
His avatar of an Asian woman was given the generic name “Jackie Morris”, plus a false Instagram account and a typical set of family and social images, all AI-generated. “She” was applying for a post as a customer acquisition specialist. Mr Moore spent just $50 to buy software which changed his voice to that of someone who might appear to be a youngish female of the relevant origin. The use of a generic name is crucial – it must be a name which, when searched on Facebook say, will throw up a large number of alternatives (John Smith is probably too obvious).
There were, apparently, 261 other applicants for the role
To begin with, his avatar, Jackie Morris, spoke to the company’s own AI avatar. Many of us have already experienced this sort of process. On two occasions recently, when I was being asked to talk to banks about the trade show industry, I was first “interviewed” by someone who gave their “name”, “position” and “location”. It took me a few minutes (me being very slow on the uptake) to realise that I was talking to an avatar. The giveaway in both cases was the interviewer precisely repeating back all of my answers to their questions. I would like to think I won’t be fooled again, but I suspect that AI is advancing so fast that this will no longer be a giveaway.
Avatars will repeat back your answers
Back to Jackie Morris. When “she” was called back for a more detailed second interview, Mr Moore set up a parallel screen and chatbot which immediately suggested answers to all the questions “she” was being asked and “she” simply spoke these answers. Despite her references and all her online presence being fake, “she” was offered the job. And 261 others were disappointed.
It is slow burn avatars that will be the problem
There are a number of takeaways. The difficulty increases if a cheat needs to create real IDs, real emails, real social media accounts and real references. That is a warning sign to employers, though it is increasingly easy to “steal” personalities with all of their attendant detritus via all forms of social media. Another is that there was nothing to be gained by the avatar getting the job – there was little loss to the company apart from red faces. But it does point the way to how easy it would be to create a transaction which would cost a business real money. The creation of a long term, slow-burning, employee avatar (in a work-from-home job) could eventually lead to that avatar being trusted when it came to paying invoices or transferring funds.
But the real takeaway for the exhibition industry is how public and financial perceptions of the trade show business have changed. And that is that we are, to a significant extent, “AI-proof”. AI is a tail-wind for us at just the time investors won’t buy businesses if they cannot assess the long-term effect of AI.
The most startling demonstration of this new reality is that in the first six months of 2026, some $5.6bn was spent on acquiring trade show assets by private equity companies, which is certainly good news. But, positive indications aside – whatever your business, be aware of job applications from avatars.