Is this the end of human-driven evaluation? Assessment in the AI era

Generative AI has stormed into the limelight this year, propelled by the groundbreaking release of ChatGPT. This game-changing technology represents a significant leap forward in the capabilities of large language models (LLMs). But it's not just generative AI that is making waves. Advancements in core internet hardware, such as augmented and virtual reality (AR/VR), have captured attention, with Apple's visionary strides in this space promising a new frontier for the internet.

In light of these developments, it begs the question: What implications does this hold for the future of assessment?

Legislation

Firstly, any assessment must adhere to DEI legislation and principles of fairness which need to be empirically demonstrated.  Respect for user privacy, for example as enshrined in the European GDPR regulations, is also fundamental.  In relation to AI, there are also emergent regulations in both the EU and US which will increasingly put guard rails around how any solutions are constructed and used.  A “Responsible AI” approach should become the expected baseline for buyers. 

With the reassurance for buyers that these issues have been managed, generative AI holds immense potential for the full range of assessment contexts. One of the quick wins lies in the scoring of assessments. We already have machine learning techniques being deployed to analyse video interviews. But the possibilities extend beyond just video responses; generative AI makes it feasible to analyse written answers, assessment centre exercises, case studies, and other types of assessments.  

There is also scope to transform how traditional psychometric constructs like personality or situational judgment are measured using an interactive and discursive approach using generative AI, instead of traditional fixed-format questionnaires. By enabling scalable analysis across multiple assessment inputs, generative AI could prove to be a game-changer, driving cost reduction and enhancing scalability. 

Beyond scoring

There is also considerable opportunity to enhance how assessments and interviews are generated. For instance, AI-driven selection of the best combinations of different modes and questions to assess any specific role.  This would accelerate a move towards mass customisation and away from one-size-fits-all, off the shelf approaches.

Generative AI also has the power to vastly improve reporting and feedback, offering individuals coherent, highly relevant insights and recommendations. This technology has the potential to consolidate multiple assessment tools and deliver concise, valuable reporting at an affordable price point.  And it does not need to be restricted to traditional formats, but could be conversational in nature providing a coaching-style experience.

Interviewing

With the application of generative AI, interviewing could also undergo a revolution. While video interviewing already allows for recorded responses to be analysed at scale, imagine the possibilities of conducting real-time interviews with generative AI connected to realistic avatars. This combination would not only enable highly scalable interview experiences but also facilitate automated scoring and radically reduce the work demands for recruitment teams.

While this approach could work well for volume hiring, more specialised roles may particularly benefit from the convergence of software and hardware. Augmented and virtual reality tools, which have been in existence for some years now, are gaining traction as the hardware as evolved. Facebook's investment in the metaverse has seen around 20 million units sold of their Quest virtual reality headsets. Now, with Apple entering the mixed reality scene with their Vision Pro product, the potential for widespread user adoption and platform transformation becomes even more significant. The HR tech analyst Josh Bersin has highlighted both on-the-job applications for high risk jobs, and the likely doubling or tripling of learning and development uses of the technology.  

Apple’s carefully planned entry with Vision Pro adds credibility to the market, and as the user base grows this opens up great opportunities for new forms of assessment when combined with generative AI. Whilst it may yet take some time as adoption of these platforms takes hold, the stage is set for efficiently creating hyper-realistic simulations that mirror real job scenarios, a long-awaited breakthrough for the assessment industry.

Reshaping assessment

In sum, generative AI is poised to help us reshape the landscape of assessment. It holds the potential to not only reduce reliance on human assessors, interviewers and coaches but also enable complex assessment simulations where content can be managed quickly and easily. As we explore the exciting possibilities, it's essential to address the barriers that may impede the rapid adoption of these technologies. However, with ongoing advancements and the convergence of software and hardware, the future of assessment stands on the cusp of a transformative era.