By Ross Heron | CEO, Australian Payroll Association
In October 2025, Mercor, an AI-powered hiring platform, closed a $350 million funding round at a $10 billion valuation. By that point it had reached $500 million in annualised revenue, with clients including Meta, OpenAI and Anthropic. It was, by any measure, one of the hottest companies in the AI economy.
By April 2026, it was fighting for survival.
A supply chain attack on a widely used open-source library exposed up to 4 terabytes of Mercor’s data including candidate profiles, Social Security numbers, passports, biometric data and recorded video interviews. Meta paused its contracts, and other major clients were reported to be reviewing their relationships. At least five class action lawsuits were filed, alleging negligence, breach of privacy and violations of consumer protection law. A company valued at $10 billion was facing an existential crisis within weeks. It was triggered not by a direct attack on its own systems, but through a single compromised library it did not control.
The incident sits outside payroll. But the question it raises sits squarely inside it.
Because the real issue is not cybersecurity. It is accountability. And that question is now arriving in payroll with force.
Across organisations, there is growing pressure to adopt AI to improve efficiency, reduce manual processing and manage compliance complexity. Software vendors are actively promoting AI capabilities across award interpretation, anomaly detection, employee query handling and system integration. But promotional activity and genuine adoption are different things, and in Australia, most payroll teams are not yet living that reality.
The question being asked, almost universally, is “What can we automate?”
It is a reasonable question. But it is not a sufficient one.
In practice, many payroll teams are still at a much earlier stage in the journey than the broader conversation implies. Recent APA survey data shows that just 13% of respondents are currently using AI within their payroll function, with the majority still in an exploratory phase.
The barriers are telling. They are not primarily financial. Instead, they reflect capability and governance challenges around one third of respondents cited a lack of understanding of AI technology, while a quarter highlighted privacy concerns.
This points to a shift in the real question payroll leaders need to be asking. For most teams, the challenge is not whether to adopt AI but how to do so with confidence, control and clear accountability.
Payroll is a compliance-critical, trust-based function. Errors have immediate, tangible consequences for employees, for organisations, and for the practitioners responsible for outcomes.
As AI becomes embedded in payroll processes, the risk expands beyond calculation accuracy. It extends into data governance, explainability and decision ownership. The question is not just whether the system produces the right number, but whether someone can account for how it got there.
That accountability does not transfer to the software vendor. It stays with the employer.
AI does not reduce responsibility. If anything, it increases it. The distance the distance between a decision and its human owner grows.
A more useful frame for payroll leaders is what we might call Accountable Payroll, an approach that treats AI adoption not as a question of what is technically possible, but what can be implemented with rigour and transparency.
Accountable Payroll rests on three principles
This shift is being reinforced by regulation, and the direction is clear.
The EU AI Act classifies employment-related AI systems as high-risk, which carries specific obligations around human oversight, transparency and documentation. While Australia is not yet subject to that framework, it signals where global expectations are heading, and multinational employers are already navigating both.
Domestically, Privacy Act obligations already require organisations to protect personal information, with reforms expected to sharpen accountability requirements further. And Fair Work obligations remain unchanged. Employers are responsible for pay outcomes regardless of which systems produced them.
The question is not whether payroll will be held to these standards. It is whether it will be ready when it is.
As AI becomes part of payroll infrastructure, there is one question that cuts through the noise
If an AI-enabled process produces an error, can you explain it, fix it, and take responsibility for it?
If the answer is unclear, the system is not ready to scale.
This is not an argument against adoption. AI presents genuine opportunities for payroll teams to reduce manual burden, improve accuracy and focus human expertise where it matters most. Those opportunities are real and should be pursued.
But payroll has always been built on precision, compliance and trust. AI does not change that foundation. It raises the stakes for protecting it.
The question for payroll leaders is not what can be automated.
It is what can be automated responsibly, and who owns the outcome when something goes wrong.
Download our Payroll AI Adoption report https://austpayroll.com.au/payroll-ai-adoption-report