Business leaders and workers need to start experimenting with AI tools “immediately” or risk being left behind, an expert on the future of work has warned.

This week, a Goldman Sachs report estimated around 300 million jobs around the world will be displaced by AI products like ChatGPT, Microsoft’s Copilot suite, and image generator Midjourney with a massive two-thirds of jobs in the US and Europe “exposed to some degree of AI automation”.

Dr Sean Gallagher, director of Swinburne University’s Centre for the New Workforce, told Information Age AI tools are a net positive for worker empowerment so long as they’re thoughtfully integrated into an organisation.

“Businesses are not ready for the freight train that is coming down the track for them,” he said.

“To take full advantage of these tools organisations need to begin experimenting with them immediately.”

It’s not so much jobs disappearing that is the major shift with the long-awaited rise of AI, according to Dr Gallagher, rather the routine, repetitive, and predictable tasks that will be automated.

Some jobs are made up of these tasks more than others, as the Goldman Sachs report points out, such as administrative and legal roles.

Likewise, elements of architecture and engineering – and yes, IT and computing – have a higher-than-average exposure to AI automation.

“Organisations need to look at the job types across their organisations and identify the tasks that could take advantage of these tools to deliver better productivity,” Dr Gallagher said.

“Do things like setting aside a group of workers which does not use a tool to do a certain task, and another group that does use the tool, and see what the real time savings look like. You need to get realistic measures of what AI can do.”

Time to experiment

The importance of this type of experimentation is not to find ways of cutting staff but rather to avoid having people create busy work for themselves, filling in the gaps between tasks that have been automated.

“Workers will create work for themselves rather than be idle,” Dr Gallagher said. “A lot of that work is ineffective and inefficient for the business but gets engrained and has business processes built around it, despite not being aligned with primary objectives.”

Tools could also be used inefficiently by people who waste time perfecting the fine details on an AI-generated PowerPoint presentation, thereby making the task take as long as it would have without the new software.

It’s up to business leaders to start creating guardrails and frameworks around AI to ensure it’s being used to “augment and elevate workers in their jobs”, Dr Gallagher told Information Age.

“Empowering workers is what I think the real value of this tool will be. There will be some job losses but many more jobs created.”

Despite only being a few months into the generative AI hype and entirely new job roles are already appearing.

This week Bloomberg reported that non-technical AI whisperers called “prompt engineers” are commanding salaries as high as $500,000 (US$335,000).

Prompts are the natural language commands we give to generative AI systems like ChatGPT.

US company Anthropic is willing to pay up to half a million dollars for a job that involves building “a library of high quality prompts or prompt chains to accomplish a variety of tasks” to help users with their AI needs.

Despite commanding a salary with a lower bound of $260,000 (US$175,000), the job’s only technical requirement is “at least basic programming skills” and the ability to write “small Python programs”.