Security and safety concerns have six in 10 managers at Australian businesses worried about near ubiquitous artificial intelligence (AI) tools, according to a new survey that found barely half have policies governing AI’s use – despite its potential to affect 1.4 million jobs by 2027.

Some 72 per cent of 318 Australian IT services decision-makers, polled in a new Datacom survey into attitudes towards AI, said their businesses are already using it.

With 86 per cent supporting employees’ use of tools like ChatGPT to support their everyday work activities, two-thirds of the holdouts expect to be using the technology within the next year or two, and a further 13 per cent admit AI is inevitable within three years.

Amidst surging usage, however, businesses still lack mature processes for controlling AI’s risks – with just 52 per cent saying they have established staff policies to manage its usage, 40 per cent having legal guidelines in place, and 39 per cent indicating they have audit assurance and governance frameworks in place.

Given near universal agreement that AI’s adoption will impact businesses’ operations and workplace structures – and earlier findings that employees aren’t getting enough training on how to use AI correctly – the lack of AI governance raises serious concerns for businesses.

“It’s important to empower employees with the right tools to support and enhance their roles, however when it comes to use of generative AI in a corporate environment a ‘proceed with caution’ approach is needed,” Datacom group chief information security officer (CISO) Karl Wright said, warning that unregulated use of such tools by employees raises problematic ethical, legal, and copyright issues.

“The use of AI needs to be carefully considered, monitored and governed with clear policies and guidelines in place to ensure the risks to businesses are minimised,” Wright said, adding that companies “need to take a long-term view towards AI, which is why laying the right foundations around its use is key to future success.”

The challenges of ChatGPT, in particular, have already led many Australian businesses into a love-hate relationship with the technology – reflected in near universal agreement that there should be legislation for AI use in the public sector.

Creating a new skills paradox

Despite broad support for AI’s adoption within Australian workplaces, the lack of mature controls reflect the risks it poses to Australian industry – with recent Pearson research, conducted for ServiceNow, finding some Australian industries are more exposed than others.

By modelling the 30,000 skills and 26,000 tasks involved in over 6,500 occupations, Pearson evaluated their exposure to AI – concluding that some 1.4 million existing Australian roles are likely to be automated using AI by 2027.

This includes 143,000 jobs in the financial services and insurance (FSI) industry, which is likely to boost productivity by 7 per cent by automating around 30 per cent of its processes.

Professional, scientific and technical services firms will see the largest relative return on AI investment – with a 5 per cent productivity gain forecast after automating just 10 per cent of their processes.

Wholesale and retail companies will see the most workers – 384,000 – affected amidst moderate productivity gains, while inefficient transportation and storage companies realise relatively low productivity benefits despite seeing 78,000 jobs affected by AI.

With around 10 per cent of Australia’s workforce set to be disrupted in coming years, it’s hardly surprising that 63 per cent of respondents to the Datacom survey believe AI implementations should be managed by specialists – yet it’s not clear where those specialists might come from.

The answer, notes Datacom Australia managing director Alex Coates, lies in retraining existing subject matter experts who can maintain a sense of “context” as they reskill to manage the legal and governance aspects of greater AI adoption.

“We see the potential to weed out more automated, routine tasks and focus that resource elsewhere,” she explained, arguing that businesses must talk not only about how AI will affect jobs, but how they can reshape those jobs to remain relevant in the AI era.

More widespread AI use will boost demand for skills like AI and machine learning specialists, and data scientists and analysts, she said, “so we need to build those skills into our existing workforce.”

“It is important that we look to the future, and implement the right skills and training opportunities now.”