With All This Talk of AI, Somebody Forgot the Human in Talent Assessment 

The waves are abuzz with talk of ChatGPT, generative AI in general, and the other AI assistants that have burst onto the scene in recent months. Applied to talent assessment, there are excited questions being asked about how this might revolutionize the way that talent assessment works, or indeed level the playing field for job applicants. 

For sure, generative AI does go a long way to levelling the playing field for job applicants, but unfortunately in totally the wrong way – and only in a way that undermines everyone and serves no one. Any job application process worth its salt is about helping the employer find the signal in the noise, to select the promising candidates from the poor candidates.  

Unfortunately, generative AI has just given us all a lot more noise. A lot more false applications, easy to produce and submit at scale, that do nothing to improve our chances of landing the right role – in fact, it even makes it worse. The wrong applicants submitting the wrong information, wasting their time and yours. Dealing with more noise burns the time, energy and resources of every recruiter – time, energy and resources that could otherwise be invested in finding and hiring the right talent. 

Refining the signal from the noise is what talent assessment has always been about. It’s worth taking a moment to remind ourselves of what talent assessment does, how, and why. Talent assessment helps identify the people who are the best match for the role and who will perform best when hired.  

It does this by working out what are the attributes (e.g., skills, strengths, competencies, capabilities, behaviors, values, motivation, experience) that are required in the role (through a process such as Cappfinity’s Futuremark), then assessing whether a given person demonstrates these attributes. And whether that person demonstrates these attributes to a greater or lesser extent, relative to every other person who applied. 

In the case of Cappfinity skills and strengths assessments, this includes an embedded focus on engagement and motivation, an emotional dimension that simply won’t be detected by a generative AI, such as ChatGPT.  

Here is what I was told when I asked ChatGPT ‘How do you feel about that?’ in reaction to a response it had given me: 

As an AI language model, I don’t have personal feelings or emotions, so I don’t have a specific sentiment about the possibility of humans re-creating and reviewing the steps involved in generating a response. My purpose is to provide helpful and accurate information to the best of my abilities based on the training I have received. 

Advanced talent assessment will use a range of approaches to elicit this information from a human, in ways that are scientifically valid, thereby building a deeper insight of the nuanced contours of human response. For advanced talent assessment, standard best practice developed over decades means that it is largely impervious to the depredations of generative AI.  

Drawing from a selection of Cappfinity assessments, this includes things like: 

  • Behavioural responses in a situational context, where there is not a right or a wrong answer, but an optimal answer that is only determined in context 
  • Time-recorded responses, where it is literally faster to respond naturally as a human, than it is to try and use an AI assistant 
  • Live interactions between humans, either synchronously or asynchronously, that include recorded verbal responses, interview answers or facilitated discussions, all then scored and evaluated by real human beings on the recruiter side. 

Talent assessment evolved as a better way of helping recruiters to find the signal in the noise. Generative AI is already creating a lot of noise and will continue to do so, but talent assessment is more than up to the task of continuing to refine the signal from that noise. 

This matters, because talent assessment has helped us make enormous advances in social justice over the last 10 years, increasing the representation of women, people from ethnic minority backgrounds, people from less advantaged backgrounds, and people with disabilities or neurodiversity. Talent assessment has done much to genuinely level the playing field of opportunity, most especially across early talent recruitment.   

The common core of all of this is the human being, the person who is at the core of the talent process. The human being, with all their subtle contours, their quirks, their brilliance and their uniquely human attributes.  

Talent assessment matters, it helps ensure we keep that human being in focus. Generative AI might be creating some noise, but talent assessment ensures we can always find the valid human signal in that noise.