AI Orientation in practice: a recap of our webinar
We recently hosted a webinar led by Ben Williams, Chief Scientist at Unseen, on how employers can assess the psychological factors that shape people’s approach to AI, alongside their current AI skills.
AI capability describes what someone knows and can do with AI today. It can be important when a role requires a particular skill from day one, but platform-specific measures may not transfer to other tools, and tests of technical skills can date quickly as technology changes.
AI Orientation offers a broader view. The webinar introduced three related areas:
Dispositions and personality: ways of thinking, feeling and behaving.
AI-specific motivation and confidence: whether someone sees value in learning, feels able to do so and is willing to invest effort.
Attitudes, values and beliefs: what someone trusts about AI, how they view its consequences and which uses they consider acceptable.
These areas can help explain why people with similar skills may respond differently to new AI tasks. What they do not do is point to one ideal AI profile, instead, the role should determine which aspects matter.
Ben then outlined how employers can put AI Orientation into practice: start with job analysis, establish the capabilities needed on entry, identify which aspects of AI Orientation matter for the role, and choose suitable ways to assess them.
How can an organisation assess AI Orientation when they have hundreds of roles?
Ben’s answer was to balance consistency with enough detail to reflect the work:
Look for clusters of roles and the requirements they share.
Apply more detail where AI use has different or more significant implications.
Avoid both a bespoke assessment for every role and treating every role as identical.
This is a pragmatic way to decide what can be assessed consistently and where a role needs closer attention, rather than a fixed formula for grouping jobs.
Can AI Orientation develop after someone is hired?
One employer asked about development over time and specifically what to do when employee attitudes change, for example if they become more cynical about AI. Ben noted that the right response depends on the organisation and why attitudes have changed. There is no single intervention that will shift every team’s views.
Personality tends to be more stable, while confidence, motivation and ethical views can change with experience and context. Development conversations can explore what is behind a person’s response. For example, someone who doubts they can learn a new tool might reflect on a time they successfully learned new technology. Someone who sees little value in AI, or objects to a particular use, needs a different discussion.
What if a candidate objects to AI in selection?
Ben shared a recent story from an event where a panel of students said they might react negatively to AI in recruitment, including on environmental grounds. Other candidates may welcome AI in selection because they expect a quicker process or believe it could reduce bias.
Employers should understand whether the concern is about the selection method, a particular tool or AI use in the job, then assess candidates against the actual role requirements using consistent evidence.
Volume-proofing the candidate experience: Key takeaways from our early careers breakfast
Applications numbers are still rising, competition is increasing and AI is changing recruitment processes faster than most employers can adapt.
These were some of the themes at our recent breakfast event, in partnership with the ISE, hosted at The Ivy in Tower Bridge.
Early careers leaders from industries across Finance, Retail, Engineering and more gathered to hear the latest findings from the ISE’s 2026 Student Recruitment Survey, explore the psychology behind the candidate experience, and compare approaches with one another. It was a fantastic opportunity to compare the quantitative data against what employers are seeing in their own processes.
Read on for the key findings, the discussions around the room and how business psychology and behavioural science can help employers respond to the latest market insights.
AI is disrupting process faster than it is being adopted
Dr Claire Tyler, Head of Insights at the ISE, presented findings from the latest Student Recruitment Survey. It drew responses from 155 employers, who received 1.7 million applications and made over 31,000 hires last year.
Employers are not managing candidate use of AI by simply adopting AI themselves. Whilst 77% are reviewing or redesigning their process as candidate use of AI climbs,, only 16% of employers use AI in recruitment today. The influence of AI is clearly extending beyond organisations actively deploying it in their recruitment process.
There’s been a significant shift in employer attitudes towards candidate use of AI, becoming more accepting that it has become part of students’ everyday lives.
64% are now comfortable with candidates using AI at some stage of the recruitment process, up from 46% last year. 82% provide guidance on how to use AI appropriately, compared with 55% in 2025.
At the same time, 92% of employers said they had encountered some form of cheating through the application and assessment process.
Discussions around the room
Where is the line on AI?
Several employers are testing and exploring using AI in recruitment rather than committing and governance is still a significant consideration for many.
There was also an interesting discussion on where automation ends and AI begins. Automation has long been part of recruitment, but now it seems the two are often being confused, with some candidates who receive automated responses assuming AI is involved. It was agreed, therefore, that the language employers use needs to be carefully considered, as well as some work from employers to increase understanding amongst candidates about the distinction between AI and automation.
Keeping assessments fair and human
Some employers have moved interviews to in-person after seeing candidates rely on AI during live conversations. Others have dropped recorded video interviews and virtual assessment centres. For both, the trade-off was cost and attendance levels. One employer reported 30% of candidates dropping out of in-person assessment centres, particularly where candidates had other offers that did not involve travel.
In a world where candidates are making many more applications than usual, it’s understood that they cannot realistically feel as strongly about every role they apply for – application volume alone does not secure the candidates employers want. Ben Williams, Chief Scientist at Unseen, took the group through the psychology behind the candidate experience and ways to improve candidate commitment throughout the process, including providing more realistic job previews, justifications of chosen assessment tools and sharing testimonials from relatable colleagues.
The next biggest theme was application volume and the impact it can have on the candidate experience.
It was found by the ISE that graduate hiring is stabilising, but that the growth remains uneven. Hiring rose by 4% this year after an 8% fall last year. The contraction has stopped, but growth is not returning across the board. When they exclude the small number of very large employers from the dataset, the average change is close to zero.
At the same time, volume of applications continues to be the key issue for employers. 73% of employers expect high application volumes to be a challenge in the next recruitment season, and 60% say managing high application numbers is a top concern for the next 5 years. Application numbers are still at a record high, with 141 per vacancy overall, rising to 172 in London.
Employers are responding to this by adjusting their processes using shorter application windows, basic screening automation and limits on how many roles a candidate can apply for. Using AI to manage volumes was low down on the list of adjustments being made, ranked 10th.
Discussions around the room
Consistent communication
After hearing the data, discussions turned to how this could be put into practice, particularly at scale.
It was agreed that automated messages to candidates can still be rich, tailored and informative, even when they’re not coming from an individual. One employer shared that they use a single recruiter to record a video with updates at each stage of the process, so candidates can feel better connected and valued by seeing the same person throughout.
The challenge of keeping candidate attendance during live events was brought up once more, and some employers shared that they have moved pre-joining webinars to pre-recorded sessions after seeing candidates drop out of live events.
Ben shared how a psychologically-informed journey design can help organisations maintain candidate attention and engagement across long lead times, rather than relying on isolated moments of communication. Unseen’s candidate and onboarding experience platform, Meet&Engage, facilitates this through staged content releases, reminder emails and notifications, structured content ordering and progressive engagement over time.
Delivering feedback at scale
Only 14% of employers in the ISE survey said they provided feedback at the initial screening stage, rising to 40% at testing. One employer shared how through utilising Unseen’s Assessment Centre platform, TopScore, they were able to easily provide quality feedback at every stage of the process and rarely saw any challenge on that feedback.
Another discussion was about the value of the data collected through assessment – one employer shared how they use assessment centre data to build a candidate profile at the end of the process, and everyone was in agreement that it would be useful to be able to join up that data with candidate strengths and improvement areas later down the line once they were in role.
How we can help
Volume-proofing the candidate experience means scaling processes without diluting the journey. Everything employers do needs to centre on being clear, relevant, human and trusted.
The topics covered at the breakfast are ones we work on with employers every day, from designing assessment journeys and feedback to keeping candidates engaged between stages and planning for volume.
If your team is facing similar pressures, we would welcome a conversation. Get in touch with us to discuss how these findings apply to your early careers programme and where we might be able to help.
A psychological analysis of why assessing for today’s AI skills is no longer enough
In a short space of time, the concept of being “good at AI” has changed significantly as the tools have rapidly evolved. For employers, building assessments to understand candidate AI proficiency can therefore feel like chasing a moving target. How can we effectively evidence the skills we need from candidates when the tools are changing so quickly?
Our latest whitepaper by Ben Williams, Chief Scientist at Unseen, introduces the concept of AI Orientation: a more enduring way to explore the motivations, confidence, values and beliefs that shape how people engage with AI; and how to turn those findings into useful assessment evidence.
Knowing what someone can do with AI today does not tell you everything about how they will approach tomorrow’s possibilities. AI Orientation considers the attitudes, motivations and personal boundaries behind that approach which allow us to understand a longer-term view of the individual.
01 / PSYCHOLOGY
The psychology behind AI use
Our research-informed approach brings together three perspectives that offer a richer picture than asking whether someone is “good with AI”; personality, motivation and beliefs & attitudes.
02 / PERSONALITY
What you’re already assessing
We explore how your existing assessments may already be providing the right kinds of evidence about some of the psychological foundations of AI Orientation.
03 / APPLICATION
Applying the approach
The whitepaper looks at how you can apply job analysis to AI-enabled work, the legal considerations and how this can be relevantly applied to selection criteria.
An example from the whitepaper
Understand why people approach AI differently
People with similar AI skills may need different support. AI Orientation explores the dispositions, motivations, attitudes, values and beliefs behind their approach. Open a statement to see how this could shape a
development conversation.
“I don’t think I can learn this.”
Confidence in learning
Willingness and self-belief are not the same.
Someone may want to use AI but doubt their ability to learn.
A supported, manageable task could be more useful than asking
for greater enthusiasm.
A useful starting point: what feels difficult, and what have
previous learning experiences shown?
“I can’t see the point.”
Motivation for AI
Find the purpose, not just the enthusiasm.
Someone does not have to find AI fascinating to have a reason
to use it. Explore whether an appropriate application could
help them achieve an outcome they value.
A useful starting point: what real problem could AI help with,
and what would the trade-offs be?
“I disagree with how it is used.”
Values and beliefs
A concern is not automatically a development need.
An objection might relate to one application, a supplier or a
settled ethical principle. Understand its scope before assuming
that more training will change it.
A useful starting point: which uses feel unacceptable, and would
any safeguards address that concern?
Illustrative development conversations from Beyond AI Literacy.
These are starting points for discussion, not an assessment or a diagnosis.
Inside the whitepaper
A more considered way to understand AI skills
Beyond AI Literacy explores how AI Orientation can inform conversations
about hiring and development, while keeping assessment connected to what
a role actually requires.
✓Look beyond current technical skills to the motivations and beliefs that shape AI use.
✓Consider how existing assessments can be complemented by structured interviews and tailored exercises.
✓Use a more personal starting point for development conversations about AI.
In this webinar, Ben Williams, Chief Scientist at Unseen, dives into our whitepaper, Beyond AI Literacy, which introduces the concept of AI Orientation.
He explores why assessing today’s AI skills is not enough, and how understanding people’s motivations, confidence and boundaries can inform hiring and development.
People responsible for hiring, assessing or developing talent as AI changes how we work.
What we discuss
How the whitepaper’s approach can help you assess AI readiness beyond current technical skills.
An example from the whitepaper
Understand why people approach AI differently
People with similar AI skills may need different support. AI Orientation
explores the dispositions, motivations, attitudes, values and beliefs
behind their approach. Open a statement to see how this could shape a
development conversation.
“I don’t think I can learn this.”
Confidence in learning
Willingness and self-belief are not the same.
Someone may want to use AI but doubt their ability to learn.
A supported, manageable task could be more useful than asking
for greater enthusiasm.
A useful starting point: what feels difficult, and what have
previous learning experiences shown?
“I can’t see the point.”
Motivation for AI
Find the purpose, not just the enthusiasm.
Someone does not have to find AI fascinating to have a reason
to use it. Explore whether an appropriate application could
help them achieve an outcome they value.
A useful starting point: what real problem could AI help with,
and what would the trade-offs be?
“I disagree with how it is used.”
Values and beliefs
A concern is not automatically a development need.
An objection might relate to one application, a supplier or a
settled ethical principle. Understand its scope before assuming
that more training will change it.
A useful starting point: which uses feel unacceptable, and would
any safeguards address that concern?
Illustrative development conversations from Beyond AI Literacy.
These are starting points for discussion, not an assessment or a diagnosis.
What you’ll take away
A more useful view of AI readiness
01
Understand what sits behind AI use
Distinguish technical capability from motivation, self-belief and
ethical boundaries, without reducing people to an enthusiasm score.
02
Build on your existing assessments
Explore what personality questionnaires already cover, and where
structured interviews and tailored exercises can add evidence.
03
Make development more personal
Identify the conversation each person needs, while keeping assessment
grounded in what their role actually requires.