Australia acts on AI incidents as insurers and investors raise pressure.

The insurance industry is reacting to the rapid growth in AI-related risks by rethinking coverage. Insurers have begun introducing new policy exclusions to avoid unquantified AI liabilities ([1]). This summer, a survey by global broker Gallagher found a 978% rise in AI-driven lawsuits between 2021 and 2025 ([2]). These lawsuits span issues from automated defamation to algorithmic discrimination. Insurers fear that traditional liability and cyber policies, which were not designed with AI in mind, could be on the hook for massive losses that haven’t been priced in. In response, industry groups have drafted standard clauses to exclude certain **“generative AI” risks, such as malicious outputs or autonomous failures, from general coverage, unless explicitly added back at higher premiums ([3]) ([4]).

The result is a widening “silent AI” gap, where many businesses’ exposures to AI-related failures are largely uninsured ([5]) ([6]). Some specialist insurers have launched dedicated AI liability products, but these are new, often costly, and targeted at larger firms ([7]) ([8]). For most companies, the immediate takeaway is that transferring AI risk to insurers will be harder and more expensive. Boards and risk officers should verify how existing policies treat AI incidents and consider additional coverage or self-insurance if needed. The surge in lawsuits also serves as a stark warning: organisations must implement rigorous risk management for AI systems, especially where they interact with consumers or critical operations, to avoid becoming part of that statistic. Effective AI governance and internal controls are becoming essential not just for compliance, but also for basic financial protection.

Artificial intelligence has rapidly moved from a niche tech topic to a core governance concern for corporate boards. A recent analysis by Morgan Stanley found that roughly two-thirds of S&P 500 companies now have at least one board director with AI expertise ([1]). But despite this progress, overall board fluency in AI remains low: The Conference Board reports that less than 3% of large-company directors have specific AI expertise ([2]) ([3]). Meanwhile, shareholders are increasingly impatient for demonstrable oversight. According to proxy research firm ISS, AI-related shareholder proposals have risen over the last five years even as other environmental and social proposals declined ([4]). These proposals have urged companies to assess AI’s impacts, improve transparency around algorithmic decisions, and ensure an executive is accountable for AI risks.

This trend shows that investors view AI governance as integral to long-term corporate value and risk management. Regulators, too, are starting to ask how companies’ leadership is managing AI. Forward-looking boards are responding by setting up dedicated AI ethics committees or expanding risk committees’ mandates to include AI. Leaders are formalising internal AI policies and oversight structures, yet a Thomson Reuters Foundation/UNESCO study found that only 13% of companies worldwide have a defined policy for human oversight of AI systems ([5]) ([6]). To meet growing stakeholder expectations, businesses should ensure they have clear governance frameworks: from inventories of AI applications to board-level reporting on AI initiatives and risks, and a single senior executive in charge of making AI both safe and effective.

Around the world, public authorities are intensifying their warnings that stronger AI governance is urgently needed. This week the United Nations High Commissioner for Human Rights, Volker Türk, cautioned that “the world is running out of time to put guardrails on artificial intelligence” ([1]). He deplored the current “ruthless race” between companies and nations to deploy ever more powerful AI systems without adequate oversight ([2]). At an international summit in the UK, Mr. Türk urged a ban on allowing machines to decide “who lives and who dies” on the battlefield ([3]), and warned that AI is already being used in ways its creators “cannot fully predict or manage” ([4]). These remarks come on the heels of similar pleas from scientific and civil society leaders for greater transparency and accountability in AI development.

The High Commissioner’s concerns echo those of national regulators and experts who fear that AI advancement is outpacing society’s ability to govern it safely. His comments add moral weight to the push for laws and standards that enforce basic principles like human oversight and risk assessment in AI. The European Union’s AI Act has already begun to take effect, setting a precedent by requiring extra safeguards for “high-risk” AI systems ([5]). Other jurisdictions, including the United States, are considering new oversight bodies and regulations to address AI’s potential harms ([6]) ([7]). For businesses, this global drumbeat means that voluntary measures may soon give way to mandatory rules. Companies that proactively align their AI practices with emerging norms, ensuring their algorithms are transparent, secure, and subject to human control, will be better prepared for the regulatory regimes now on the horizon.