The World Economic Forum has indicated that 85 million jobs may be displaced by the evolution in the division of work between humans and machines by 2025.
However, it has also estimated that 97 million new roles may emerge in areas such as AI ethics, human AI interface, and reviewing of AI output.
Forbes, the influential US finance magazine, has cited factory workers, couriers, investment analysts, customer service operatives and security guards as being most at risk from AI.
PwC director of people and organisation consulting Laoise Mullane says the recent accelerated adoption of generative AI, such as ChatGPT, has the potential to have a far broader impact on the workforce than other forms of AI and automation.
“It has the potential to significantly disrupt and displace a large share of tasks performed by workers,” she says.
“Jobs that focus on content creation, such as code, text, audio, video, images, may start to be impacted first – programming, coding, copywriting, customer service to name a few.
“Roles will shift to have a bigger focus to reviewing, refining and assuring the quality of the outputs of generative AI systems.
“While AI allows you to build better models to analyse data and automate decisions it still lacks the business understanding to turn these results into meaningful insights and actions.”
An AI tool such as Microsoft’s Co-pilot, can help a programmer get started, but if someone can’t make sense of the code it produces, then how will they determine the quality of the work?
— Robert O’Connor, a lecturer in computing at South East Technological University
Robert O’Connor, a lecturer in computing at South East Technological University (SETU), says identifying what jobs are in danger is “a huge question and probably can’t be answered quickly”.
“In my own field, computer science, we can see how AI tools can be used to write code,” he says. “Does this mean the death of programmers? I don’t think so.
“An AI tool such as Microsoft’s Co-pilot, can help a programmer get started, but if someone can’t make sense of the code it produces, then how will they determine the quality of the work?
“Ideally, an AI might take you 70 per cent of the way there, but the remaining 30 per cent is coded by a human – and that will usually be the trickiest 30 per cent.
“Hopefully, it will help us write better software overall because it frees up our time to work on more complex problems rather than focusing on the mundane tasks.”
