Healthy AI Adoption Rates… but is Your Job Description Keeping Up?

Hello Humans! And welcome to the weekly podcast that aims to cover a key AI issue relevant to HR in around five minutes. This week I want to talk about the impact of AI on job descriptions. We used to say things like “what use cases are there for AI?

” You don’t hear that expression anymore because AI has developed so quickly it seems there’s a use case for every bit of it now in the workplace. The findings from the Working in Ireland Survey published this week suggest that the island of Ireland adoption rate of GenAI at 28.4%, is comfortably above the European Union average. Since early 2023, Legal Island has been tracking the adoption of GenAI by HR and this has been slow but steady.

Our latest poll of HR professionals in June of this year indicated for the first time a majority of them (55%) were using GenAI at least once a week for some HR tasks. But when did you last update a job description because AI changed the job? An employee might now be supervising AI agents, building a basic app or producing work that previously required outside specialists. Yet their job description still describes the role they were recruited into three years ago.

For HR, that raises a question: how do we give people clarity about their jobs when the possibilities keep changing? Job descriptions have always needed some flexibility. But AI is putting that flexibility under greater pressure. Employees can now use AI to create software without writing the code themselves.

Tools such as GPT-6 Astra released just this week offer new possibilities in 3D modelling, web-app creation and game development. That doesn’t make every employee a qualified developer or designer. Producing something and knowing whether it is fit for purpose are different skills. But it does mean that the boundaries of a role can move quickly.

And there’s another shift. Some employees are beginning to supervise AI agents as well as people. By agents, I mean software that can carry out a sequence of tasks towards a goal. Someone has to brief those agents, set boundaries, check their work and intervene when things go wrong.

An employee might have no direct reports but still carry substantial responsibility for work performed by AI. Shouldn’t their job description recognise that? So how should HR respond? I’d suggest three things.

First , give the job description a stable core. Be clear about why the role exists, its main responsibilities, the outcomes expected and the employee’s authority. Those foundations should survive a change of software. Second, maintain a current work plan alongside it.

This can capture agreed AI-assisted tasks, responsibility for supervising agents, experiments and learning goals. You don’t need every new tool written into the job description. You do need a shared understanding of what the employee is expected to do. Third, agree how changes will be reviewed.

When will the employee and manager discuss them? What training and time will be available? When might additional responsibility justify a review of workload, recognition or grading? Without that conversation, flexibility can become a polite word for an ever-expanding job.

And “other duties as required” starts doing an extraordinary amount of heavy lifting. There’s a fairness question here too. If AI saves someone five hours, do we immediately fill those hours with more tasks? Or do we discuss opportunities to improve quality, learn and tackle neglected work?

Being able to do more doesn’t automatically settle what someone should be expected to do. Here’s one practical step. Choose a role. Sit down with the employee and their manager.

Ask: what has AI removed, what has it changed, and what has it added? Then agree what needs updating, and what support is required. A job description cannot predict every opportunity AI will create. But it can give people clear responsibilities and a fair process for adapting them.

Does yours? As always thanks for listening. Until next week. Bye for now!