New Zealand’s Accident Compensation Corporation (ACC) is facing increased scrutiny from claimant advocates and sector observers over its approach to exiting long-term claimants and its use of artificial intelligence to help identify cases for review. The developments are drawing attention from insurers and intermediaries to how the statutory scheme is balancing financial sustainability, rehabilitation, and data-driven decision-making.
For former aircraft software systems engineer Jacob Hamlin, life since a concussion five years ago is broken into short stretches of activity. Each day, the 40-year-old father measures out roughly four “productive hours” before cognitive fatigue, balance issues, and confusion set in. The symptoms began after he slipped on an icy path, hit his head on a square post, and was diagnosed with a concussion. Over time, chronic fatigue, brain fog, and balance problems meant he could no longer remain in his role.
According to RNZ’s report, ACC initially paid weekly compensation, but in January, the payments stopped. Hamlin received a letter stating ACC believed he should have recovered by that point and that a pre-existing condition must be responsible for his ongoing symptoms. “I received a letter saying they think I should be better by now, ending support, and that there must be a pre-existing condition causing my issues,” he said.
ACC said a neurologist reviewed Hamlin’s case and concluded that his ongoing problems related to pre-existing conditions, including ADHD, migraines, and a possible sleep disorder, rather than the covered head injury. Hamlin disputes that assessment. He said those conditions did not prevent him from holding a demanding position prior to the accident and that the loss of income has placed his family under considerable financial pressure. “When I do too much my brain stops working. I lose my balance. I get really confused. If I do something physical like mow the lawns, I can do an hour, then have to take an hour break,” Hamlin said.
After ACC cut his payments, Hamlin obtained a second clinical opinion. That neuropsychological assessment found his current symptoms were attributable to the concussion and suggested they may be permanent. An independent reviewer overturned ACC’s decision about two months ago, but he continued to wait for compensation and backpay to resume. “We’ve spent all our savings… we’re just holding on until they pay us,” Hamlin said.
Hamlin added that private health insurance funded specialist reports that were instrumental in his review outcome. “If we didn’t have health insurance to get those specialist reports showing the proof of my injury, we probably wouldn’t have won our case,” he said. Hamlin said his experience left him concerned that “the system” felt “more adversarial than rehabilitative” and that “they’re just denying everyone and seeing what sticks.”
Hamlin’s case sits within a wider shift in ACC’s handling of long-term claims. In the year to June, more than 8,000 long-term claimants were “exited” from the scheme, with ACC planning to exit more by June next year as it works to reduce the size of the long-term claims pool of around 25,000 people. Newly released figures obtained under the Official Information Act show that while ACC reported 8,741 people as having left the long-term pool, the total number of exits was 10,682 once re-entries are included. Of those, 1,941 people were exited and later reclassified as long-term claimants again.
Outcomes data indicates that only a minority of those leaving the pool returned to their original employment or were formally retrained. Thirteen percent were recorded as going back to their pre-injury role, and 3.6% were retrained for alternative work. For 5,333 of the 10,682 people who exited, ACC reported that it could not “accurately determine” the primary reason they left the long-term pool. The agency said some of those exits were likely related to vocational independence or returns to pre-injury work, but that a full breakdown would require a manual review of claim files.
Over the past five years, 42,733 people have left the long-term claim pool as ACC has confronted higher claim volumes, longer recovery times, and pressure on levy settings and liabilities. ACC Minister Scott Simpson has directed the corporation to reduce the number of people receiving weekly compensation for more than 12 months. ACC said decisions are made case by case, based on clinical and vocational assessments. However, claimant advocates report that many of the people they support are being exited on “causation” grounds, where ACC concludes that an ongoing incapacity is no longer due to the original covered injury, but to other physical or mental health conditions.
Wellington lawyer and advocate Warren Forster, who obtained exit data from ACC, describes the pattern he sees as evidence of cost and liability pressure shaping long-term decision-making. “There is absolutely a pattern of systemic exit. There’s no doubt whatsoever that ACC is exiting long-term claimants at a scale that’s unprecedented,” he said, as reported by RNZ. Forster said a significant proportion of his clients exited because ACC determines that their current condition is no longer caused by the covered injury. He argues that this can overlook complex or evolving diagnoses. “People aren’t off work for four years because of a sprain. Something else is going on – a tendon tear, a disc prolapse, mental distress – and none of that analysis is being done,” he said.
Advocate Daniel Wood, who represents claimants at review, said that while the legislation is clear, implementation can be inconsistent and review timeframes long. He estimates that about 30 of his clients have been exited from the long-term pool in the last four months, with many choosing to challenge those decisions through independent review. “They have no option but to challenge the decision through an independent review, which can take at least six to eight months –and that’s optimistic,” he said. During that period, he added, households can face “significant financial hardship,” with some relationship breakdowns and self-harm incidents reported.
ACC chief executive Megan Main said in a written statement that ACC “considered each client’s entitlements on a case-by-case basis using clinical notes and evidence.” She added: “If a client feels that we have made the wrong decision on their claim or entitlement, we have a robust independent review process in place.” Main has also highlighted ACC’s statutory duty to levy payers. “In addition to our responsibility to support people to be rehabilitated, to recover, we also have a responsibility to all New Zealanders who pay their levies to make sure that ACC is only funding injury related treatment,” she said in a recent radio interview.
ACC’s use of AI to help identify long-term claimants for further review is adding another layer to the discussion. The organisation has confirmed it is using AI models to scan long-term claims information and flag cases where the pattern of absence or diagnosis may warrant closer attention from staff. “ACC is using AI to help identify cases where the client may benefit from a further review by staff member to help ensure that they’re receiving the right support,” said ACC deputy chief executive Michael Frampton. According to ACC, the tools generate recommendations for next steps, but human staff make final decisions on entitlements and exits. Frampton said ACC’s priority is to assist injured people to regain “independence or adapt to new circumstances” after serious injury. “It’s important that we manage the scheme sustainably so it’s there for Kiwis when they need it, now and in the future,” he said.
Forster is critical of the way AI is being deployed in this context. “Nothing in ACC’s proposals actually looks at rehabilitation. It’s all about ‘how can we ramp up exits, how can we get more people coming off the scheme’, and that’s the antithesis of the statutory purpose. It’s the antithesis of why we have ACC in New Zealand,” he said. Forster argues that while AI may be able to highlight apparent anomalies, it cannot diagnose underlying causes or substitute for detailed clinical review. “The AI will tell you, ‘person X has a sprain, they’ve been off work for six years’ or whatever the case might be. AI can tell you that this person has a problem, but AI is not going to tell you what that problem is. What we have is AI identifying someone who ACC can exit,” he said.
