ANZ firms stay upbeat on AI despite governance fears

DLA Piper has published research showing that businesses in Australia and New Zealand remain confident in artificial intelligence despite concerns about governance and accountability. The study found overall confidence in AI across the region stood at 75 per cent.

The survey of 975 senior decision-makers across 13 markets also identified gaps in supplier oversight, contractual protections and internal governance, suggesting enthusiasm for AI adoption is outpacing some organisations' control frameworks.

Among respondents in Australia and New Zealand, 69 per cent said others in their AI value chain were trying to pass on liability. Another 60 per cent said their commercial strategy relied on AI suppliers over whom they had no oversight or audit authority.

The findings add to evidence that companies in the region are pressing ahead with AI investment while facing growing scrutiny over how decisions are made, documented and assigned across supply chains. They also suggest boards are being drawn more directly into issues once handled mainly by technology teams.

Nick Valentine, Partner and Head of Technology and Data at DLA Piper in New Zealand, said the data showed a widening divide between ambition and readiness. "AI has moved away from the innovation team and into the boardroom, changing the risk profile," Valentine said. "Organisations are no longer only asking whether AI can create value - they are questioning whether they can explain it, control it, contract for it, and stand behind the decisions it supports. The data suggests many organisations are building critical operations around suppliers they may not be able to oversee, audit or hold fully accountable. That is a governance issue, a procurement issue and, increasingly, a board-level accountability issue."

Supplier pressure

Supplier transparency emerged as one of the region's sharper concerns. Eighty per cent of AI adopters in Australia and New Zealand said suppliers could not meet explainability and transparency requirements, nine percentage points above the global average in the study.

That concern appears to be feeding into contract reviews. The research found 36 per cent of AI adopters in the region were reviewing commercial contracts and strategic partnership deals to limit AI liability, compared with 29 per cent globally.

Separate responses showed 32 per cent of regional adopters viewed contractual disagreements and failures as a primary risk they were preparing for, compared with 25 per cent globally. Cloud outages or compute failures, cybersecurity attacks, IP theft, and IP or copyright infringement claims were also cited at higher rates than the global average.

Joel Cox, Partner and Head of Technology at DLA Piper in Australia, said AI planning had become more developed among clients. "Boards and executive teams are looking for ways to transform key parts of their business with the use of AI, and they are focused on building the necessary infrastructure, team and culture for that. The harder questions are now coming into focus: how much control do they really have over the AI supply chain, what data can they practically and legally leverage to build their own AI models and sovereign capability, what are the costs involved and what alternative funding is available, and are their technology governance settings strong enough to support AI adoption at scale."

Governance gaps

The study indicated businesses see internal governance as an area needing more structure. Fifty-one per cent of respondents said their organisation's AI governance would be more effective with skilled or specialist oversight, while 44 per cent wanted greater accountability through one team or individual responsible for setting and overseeing AI activities.

Responses on employee use of AI also pointed to uneven preparedness. Sixty-nine per cent of respondents said their organisation took a hard line on employee AI use, limiting staff to a narrow set of sanctioned tools and use cases, while 16 per cent said their organisation had no clear position and employees largely filled the gaps themselves.

Workplace risk featured elsewhere in the findings. More than two-thirds of respondents said their organisation should be concerned about a wave of unfair dismissal cases related to AI-driven workforce changes, and 31 per cent said a major dispute or class action tied to the use of AI in workforce management or hiring was likely.

Regulatory backdrop

The research was published as policymakers in both countries continue to refine their approach to AI governance. In Australia, the government has established an Office of AI and outlined plans for a new national AI standard, including mandatory requirements for large data centres covering energy and water use.

In New Zealand, official guidance has continued to emphasise governance, security, procurement, skills, transparency, bias, privacy and human accountability in the use of AI.

The broader regional picture also reflected concern about concentration in the AI market. Seventy-nine per cent of respondents said they were concerned about control over essential parts of the AI value chain being concentrated in a small number of dominant operators, while 61 per cent said regulatory controls on compute and data location had forced them to rethink or delay AI projects in the previous 12 months.

Daniel Street, Partner, Litigation and Regulatory, DLA Piper in New Zealand, said disputes were likely to follow rapid adoption. "The next wave of AI disputes may come from missed expectations - procurement promises, data rights, performance claims, integration failures, outages, customer harm and unclear recourse when a supplier cannot explain how the system works," Street said. "The research shows ANZ businesses are already reviewing contracts to limit liability. That is sensible, but contracts need to be backed by practical governance, testing, documentation and escalation processes."

Valentine said the organisations that gain most from AI will be those that treat governance as part of commercial execution. "Done properly, AI governance gives boards, customers, regulators and investors confidence that the organisation can scale AI responsibly," he said. "The point of difference will be execution - with clear ownership, supplier discipline and resilience built in from the start."