Businesses sleepwalking toward AI oversight cliff

A widening AI governance gap is emerging as organisations accelerate investment in artificial intelligence without putting adequate oversight mechanisms in place, according to new research.

The global research, including Australia, draws on an AI-enabled review of more than 100 annual reports from multinational firms, combined with two international surveys of over 850 senior business leaders. While executives continue to prioritise productivity gains and cost efficiencies, many may be underestimating the risks, with the report findings suggesting businesses could be “sleepwalking” into governance failures.

More than half of Australian business leaders (57% vs 62% globally) expect to increase their AI investments over the next year. Their motivations are consistent with global trends: boosting productivity and efficiency (68% vs 61% globally) and managing costs (53% vs 49% globally). Most executives now see AI as central to growth, with 58% in Australia (vs 59% globally) considering the technology crucial to their future success.

However, governance practices are lagging. Fewer than a quarter of Australian organisations (23% vs 24% globally) currently operate an AI governance programme. Among larger enterprises, this rises to just over two fifths (43% vs 34% globally), but still represents a significant gap.

Susan Taylor Martin, CEO of BSI, said “The business community is steadily building up its understanding of the enormous potential of AI, but the governance gap is concerning and must be addressed. While it can be a force for good, AI will not be a panacea for sluggish growth, low productivity and high costs without strategic oversight and clear guardrails – and indeed without this being in place, new risks to businesses could emerge.”

Despite strong interest in AI transparency, reported by 81% of Australian respondents, among the highest globally, many organisations lack confidence in managing foundational risks. Australian leaders reported lower-than-average capabilities when it comes to assessing or mitigating AI bias (36% vs 48% globally), explaining data sources transparently (38% vs 51% globally), and conducting regular audits of AI systems (36% vs 53% globally).

Formal controls remain inconsistent. Two fifths (40% vs 47% globally) said AI use is governed by structured processes, up from 15% earlier in 2025. Yet only one third (32% vs 34% globally) use voluntary practice codes, and just 26% (vs 24% globally) monitor employee use of AI tools. Slightly more (38% vs 30% globally) have risk assessment and mitigation procedures. Notably, 36% of Australian firms (vs 22% globally) have chosen to restrict staff from using unauthorised AI platforms.

Charlene Loo, Managing Director, BSI Australia and New Zealand, said “Divergence in approaches between organisations and markets creates real risks of harmful applications. Overconfidence, coupled with fragmented and inconsistent governance approaches, risks leaving many organisations vulnerable to avoidable failures and reputational damage. It’s imperative that Australian businesses move beyond reactive compliance to proactive, comprehensive AI governance.”

Limited planning for errors and disruption

The research also highlights a lack of structured response planning. Only 30% of Australian businesses (vs 32% globally) have systems to track AI errors or flag inaccuracies, and just 28% (vs 29% globally) have defined incident management procedures. One in six (15% vs 18% globally) believe their operations could not continue if generative AI access were disrupted.

Despite substantial investment, value realisation remains mixed. Nearly two fifths (38% vs 43% globally) report diverting resources from other strategic projects to support AI efforts. Yet just 21% (vs 29% globally) have systems to prevent duplication of AI deployment across business units. Australian companies are also less likely to report clear time (55% vs 67% globally) and cost savings (49% vs 60% globally). At the same time, 57% said AI adoption is linked to headcount reduction—a higher rate than the global average of 50%.

Human capability and oversight lags adoption

Analysis of annual reports shows the term automation appears nearly seven times more often than references to upskilling, training, or education, suggesting human development remains secondary to technology deployment.

Confidence in workforce readiness varies. Just over half of Australian leaders (55% vs 56% globally) believe entry-level employees have the necessary AI skills, and 51% (vs 57% globally) feel their organisation as a whole is adequately prepared. Fewer (45% vs 55% globally) are confident their teams can analyse and evaluate generative AI outputs effectively.

Concerns are more pronounced in Australia than elsewhere regarding decision-making impacts. Forty-three percent (43% vs 19% globally) expect AI to weaken employee decision-making abilities, and only 36% believe AI will enhance them (vs 65% globally). This anxiety is sharper for junior staff, where 38% (vs 21% globally) worry about deteriorating judgement.

Looking ahead, 40% (vs 18% globally) believe the next generation of junior employees will see further erosion in critical reasoning skills due to AI reliance.

Training remains a critical gap. Just 28% (vs 34% globally) have structured learning and development programmes. While a higher proportion (60% vs 64% globally) have received training focused on safe or secure AI use, the report suggests this may be reactive risk management rather than strategic capability-building.

“Ultimately, investment is only the start. Successful oversight of AI tools and how they are being used is critical, as is considering the impact on people, whether that be your employees’ careers or a client’s right to privacy,” Taylor Martin emphasised. “Harnessing AI successfully requires not just enthusiasm for the potential of what AI can do, but also recognition of where guardrails are needed and how to ensure these are in place.”

In reimagining healthcare across the entire patient journey, Health Industry Hub^TM is the only one-stop-hub uniting the diversity of the Pharma, MedTech, Diagnostics & Biotech sectors to inspire meaningful change.

The Health Industry Hub^TM content is copyright protected. Access is available under individual user licenses. Please click here to subscribe and visit T&Cs here.