The hiring problem used to be described as finding the right words. Generative AI has made those words much easier to produce. A candidate can now turn a rough account of their experience into an assured application in minutes. The harder question is what an employer should believe when almost everything sounds assured.
That tension appears in the CIPD and Omni RMS’s Resourcing and talent planning report 2026, published in September. Among organisations that had attempted recruitment, 27% reported that candidates mostly or always made what the employers considered excessive use of generative AI. The finding records recruiters’ judgements, not an independently verified rate of dishonesty. Its significance is the uncertainty it reveals.
A careers fair offers a useful response to that uncertainty, provided a conversation is used well. At today’s London Careers Fair, as at any recruitment event, the opportunity is to move from general claims towards specific work. What did the candidate actually do? What does the vacancy actually require? Where might the two meet?
The cost of sounding the same
Consider two hypothetical applications. One says the candidate is an effective problem-solver with excellent stakeholder skills. The other explains how they coordinated two teams with conflicting deadlines, agreed a workable handover and kept the customer informed. Both can be written elegantly. Only the second begins to give an assessor something to examine.
AI can help a person express that second account, particularly if they struggle with writing or are applying in a language they learned later. It can also manufacture a convincing version of work they never did. Those possibilities make the tool itself an inadequate verdict on the applicant.
The useful dividing line is the reliability of the evidence. Can the candidate explain their decisions? Can they distinguish their contribution from the team’s? Can they describe what went wrong, what they learned and what they would change? A credible answer need not be fluent or dramatic. It needs to belong to the person giving it.
Nor is the employer exempt from scrutiny. A beautifully worded advert may leave unanswered the questions that determine whether someone can accept the job: hours, location, training, pay and the selection timetable. A recruitment conversation should make both sides more accountable for their claims.
Human contact is useful, but it is not automatically fair
It would be a mistake to treat an informal chat as a flawless detector of ability. Confidence, accent, familiarity and shared interests can affect an impression without establishing whether someone can perform a role. A busy room can also be demanding for a candidate who needs more time or a different way to communicate.
The CIPD’s September 2026 guidance on selection methods recommends structured interviews with consistent questions and objective scoring, and assessments that reflect relevant work. A careers-fair conversation can help identify a potential match; it should feed into a clear selection process rather than become an unrecorded substitute for one.
For an employer, that means recording a relevant fact instead of a vague impression. “Explained a stock discrepancy and the checks used to resolve it” is useful. “Good personality” is much harder to defend. Give candidates the same route to apply and explain how any assessment will work.
For a candidate, it means the encounter is not a performance of effortless confidence. You can bring notes, take a moment to think and ask for a question to be clarified. The objective is mutual understanding. If the next stage requires an adjustment, establish how to request it through the employer’s process.
Better questions than asking whether AI is used
A yes-or-no question about recruitment technology rarely tells you enough. Software might store applications, extract information from a CV, arrange interviews or assist an assessment. Those are different activities. The important questions concern the decision being made and the evidence used to make it.
Ask what the first stage assesses, which experience matters most and whether a work sample is involved. Ask where candidates can find the policy on AI assistance. If a system or task is inaccessible, ask for the adjustment route. These questions are specific enough for an employer to answer and relevant enough to improve your preparation.
The government’s Responsible AI in Recruitment guidance recognises potential efficiency benefits alongside risks such as bias and digital exclusion. Although published in 2024, its central assurance question remains useful: what evidence supports a supplier’s claims about a tool? Employers should bring that same discipline to new products instead of equating automation with objectivity.
Make the conversation a bridge to evidence
Before an event, choose two examples you can explain without a script: a problem you solved and something you learned. Read the employer’s available information. During the conversation, connect one example to a genuine requirement and ask what additional evidence would help the hiring team assess you.
There is practical precedent for reducing the distance between an applicant and a person. The CIPD’s guide to recruiting young people and non-graduates describes employer examples involving shorter forms and conversations following college careers talks. It is a modest but useful finding: alternative routes can reveal candidates who are poorly served by a long written introduction. It is not evidence that every face-to-face interaction leads to a job.
The London fair runs from 10am to 2pm at Hornsey Town Hall on 6 November, with free entry for candidates. Whether you attend this event or another, leave each useful exchange with the application route and the next step. A recruiter remembering you is pleasant; a recruiter knowing what you can do is more valuable.
As the surface quality of applications becomes easier to imitate, the premium on substance should rise. Careers fairs can serve that need when they give people a way to explain their work and employers a reason to examine it carefully. The conversation earns its place by making the eventual decision better informed.
Sources
- CIPD and Omni RMS, Resourcing and talent planning report 2026 (22 September 2026).
- DSIT, Responsible AI in Recruitment (25 March 2024).
- CIPD, Selection methods (22 September 2026).
- CIPD, Recruiting young people and non-graduates (2026).
- The UK Careers Fair, London Careers Fair (6 November 2026 event).


