Published on the 30/10/2024 | Written by Heather Wright

Accelerated career progress, if adoption can be achieved…

Generative AI will reshape jobs, helping with a third of entry-level tasks by next year, according to a new report, but the results of local pilot highlight technical, cultural and capability challenges still to be overcome.

A Capgemini survey of 1,500 leaders and managers and 1,000 entry-level employees across 15 countries including Australia, found the use of the technology could free up some staff to focus on more strategic tasks, decision making and team management.

“IT might see a move from project management to niche areas such as AI strategy or data science.”

Employees in the survey predicted GenAI would facilitate a third of entry-level tasks over the next 12 months. Half of the leaders and managers said that change will accelerate career progress for entry-level workers.

At a manager level, 51 percent of leaders and managers said advancements in GenAI would mean within three years manager-level positions would evolve towards specialisation, or move into top-tier strategic roles.

“For instance, HR roles may transition from HR generalists to talent analytics specialists or employee experience designers, while IT might see a move from project management to niche areas such as AI strategy or data science.”

Roles would become more strategic, focused on AI-enhanced decision making and performing tasks that require a high level of emotional intelligence.

The report comes as Australia’s Digital Transformation Agency notes ‘adoption challenges’ experienced during its six-month $1.2 million whole-of-government trial of Microsoft Copilot.

More than 7,600 employees across dozens of agencies were involved in the trial.

It found the technology could transform its productivity and ways of working, but agencies needed to carefully weigh potential benefits of efficiency and quality improvements against the costs, risks and suitability of generative AI to meet their agency’s needs.

The report notes back-and-forth about the implications of GenAI tools on attracting newcomers and entry level workers to the APS, and how they will transform positions.

Lucy Poole, Digital Transformation Agency general manager of strategy, planning and performance, says providing innovative tools ensures the APS remains a competitive workplace, and tools which improve the efficiency and quality of some types of work can go a long way in retaining new talent.

“This cannot be outweighed by losing key skills around the processing and synthesising concepts and knowledge of the work undertaken by an agency,” she added. “These tools should instead enhance strategic thinking and making connections between disparate pieces of work undertaken by government.”

While most workers were ‘satisfied’ with Copilot and the majority wanted to continue using it, an evaluation report notes a concerted effort to address technical, cultural and capability challenges and improve usage is required for generative AI.

Among the positives, participants felt there was a marked improvement in the speed of wrapping up tasks (69 percent), with 61 percent believing the technology enhanced the quality of work output.

Poole says participants saved on average one hour a day in note taking and administrative duties.

But feedback also pointed to the need to spend more time reviewing AI-generated content.

Poole says the DTA is ‘highly aware of the realities of bias creeping into these services’.

“AI models might replicate these biases from their training data, resulting in misleading or unfair outputs, insights or recommendations,” she says.

“That is why we continue to reiterate the importance of keeping a human in the loop.”

Despite the apparently positive sentiment, use of Copilot throughout was only ‘moderate’. Only one-third of those in the APS pilot used Copilot daily, with Teams and Word used by 70 percent, mainly for summarising and rewriting content. Poor Excel functionality and Outlook access issues hampered use in those apps.

“Other generative AI tools may be more effective at meeting users’ needs in reviewing or writing code, generating images or searching research databases,” the DTA report says.

It notes the need to tailored training and propagation of high-value use cases to drive adoption, noting that while usage was consistent across classifications and job families, specific use cases varied. For example a higher level of SES and executive level staff used meeting summarisation features.

Those findings echo those in the CapGemini report, which also flagged challenges with slow adoption of the technology – just 15 percent of leaders and managers and 20 percent of employees are using GenAI tools daily. That highlights what CapGemini says is a ‘significant’ gap between the potential of the technology and its actual usage.

It says low confidence around accuracy, logical soundness, security and respect for IP/copyright and data privacy are a key hurdle for GenAI adoption.

Complexity in integrating GenAI tools into existing workflows and a lack of skills to drive the output from GenAI effectively are further hampering adoption. Only 16 percent of employees said they receive ample support from their organisation to develop the necessary skills for GenAI, and just 46 percent of leaders and managers have had formal training.

The APS pilot results noted that just 47 percent of participants who received one form of training were confident in their ability to use Copilot. That figure rose to 75 percent for those who received three or more forms of training.

At the Gartner IT Symposium/Xpo 2024 on the Gold Coast earlier this year, Gartner flag another big GenAI challenge for companies, warning that cost was as big an AI risk as security.

Mary Mesaglio, Gartner distinguished VP analyst, said it was really easy to waste money with GenAI.

“You want AI adoption? You’re going to pay for it,” she noted, with vendors already raising prices by as much as 30 percent as they incorporate GenAI in offerings, and confusion over GenAI bills likely to mirror those of the early days of cloud use.

Gartner research has suggested nearly half of CIOs don’t believe AI has met ROI expectations.

That’s in contrast, however to a recent survey by McKinsey which showed more than 75 percent of those who have deployed gen AI at scale say its met, or exceeded, their expectations. McKinsey, however, notes that while companies are experimenting with multiple use cases for the technology, even the fastest movers have only rolled out one or two full-scale gen AI offerings.