Tego plans standalone AI cover as exclusions tighten

Tego Insurance is examining the development of what it says could be Australia's first dedicated artificial intelligence insurance product, as conventional policies move towards tighter AI exclusions.

The Sydney-based insurer is in discussions with a range of insurance partners about a standalone policy. The proposed product would address losses linked to the use of AI tools, models and automation in professional settings.

Tego focuses on professional risk in the healthcare sector. It currently does not exclude AI-related exposures from its products.

"A clear inflection point is approaching," said Eric Lowenstein, CEO, Tego Insurance. "AI exposures are evolving faster than traditional insurance products can keep up with. It won't be long before we start seeing broad AI exclusions written directly into professional indemnity, malpractice and liability policies. The market is already signalling where this is heading."

Lowenstein said early conversations with insurers centre on a policy that responds directly to AI-related claims and systemic risks. He said retrofitting AI into existing wordings would create long-term strain on traditional products.

He said any new cover would need to reflect specific AI behaviours, failure modes and systemic characteristics. It would differ from professional indemnity and liability policies that were built around human decision-making.

Rising AI exclusions

Insurers across global markets are examining AI-related exposures. Many are adjusting policy language and considering broad exclusions for certain classes of AI risk.

AI tools are now embedded in decision-making processes in multiple industries. This includes healthcare, financial services, logistics and professional services.

AI can influence or make decisions in areas that were previously controlled by individual professionals. These changes raise questions over how traditional covers respond to errors, failures or regulatory breaches linked to AI involvement.

"Insurers are already exploring limitations for AI risk. That's a clear sign of what's coming next," said Lowenstein. "The industry needs to get ahead of this now rather than wait for gaps to appear after a claim."

Healthcare at the front line

Tego identified healthcare as one of the first sectors where gaps in AI cover are likely to emerge. Hospitals and clinics are adopting AI in clinical decision support, imaging analysis, triage, remote monitoring and administration.

Most existing medical malpractice and professional indemnity policies are built around human error. Cover for AI-related risk in healthcare is still forming in many markets.

Medical Malpractice and Professional Indemnity policies typically respond when a clinician makes an error. They are less clear when an AI system assists with a decision or makes an automated recommendation.

There is also uncertainty around exposure for breaches of emerging AI regulations. These regulations may set rules for how clinicians and providers use AI within clinical workflows.

Other insurance lines may pick up some AI-related events. Product Liability policies may respond if an AI-enabled device or software causes harm. Technology Errors & Omissions cover may respond for developers if algorithms are faulty.

Directors and Officers insurance is under scrutiny as regulators monitor AI adoption in healthcare. Boards may face claims if oversight of AI systems is found to be inadequate and patients are harmed.

Clinical Trials Insurance may respond if AI is used in drug discovery or experimental treatments. Cyber Insurance may respond to AI-related breaches, system manipulation and data incidents, although some policies are introducing AI-specific exclusions.

Tego warned that clinicians and healthcare organisations could face significant exposure if traditional medical indemnity and entity liability policies introduce broad AI-driven error exclusions. That would leave professionals with limited options if AI involvement contributed to a claim.

Examples of AI risk

Insurers are already modelling different AI-related loss scenarios. These include situations in professional indemnity and medical malpractice, healthcare entity liability, and property and cyber lines.

In professional indemnity and medical malpractice, examples include misdiagnoses that are influenced by AI clinical decision tools. They also include incorrect radiology or pathology interpretations from imaging models, treatment delays from AI triage or scheduling systems, and medication errors linked to algorithm-generated prescribing pathways.

In healthcare entity liability, insurers are examining systemic failures linked to AI-driven rostering, patient management or workflow tools. They are also looking at privacy breaches or data handling errors caused by AI automation, and vicarious liability claims if staff use AI assistants in their work.

In property and cyber lines, examples include physical damage from malfunctioning AI-controlled medical equipment. Other scenarios include network outages or corrupted data linked to hospital systems that use large language models, and AI-generated coding errors in critical software.

Parallels with cyber cover

Tego said the likely evolution of AI insurance resembles the past development of stand-alone cyber insurance. Cyber risk was once an endorsement or clause attached to broader policies. It later shifted into dedicated products.

The company expects AI insurance to follow a similar path as AI becomes embedded across clinical practice, administration and broader enterprise decision-making.

Lowenstein said the first substantial AI insurance products could appear within the next 12 to 24 months. He said increasing AI adoption across healthcare operations and decision flows would drive this shift.

"We see AI underpinning whole industries as it becomes embedded across operations and decision-making," he said.

"Insurance frameworks will need to keep pace. Traditional products simply won't be able to absorb AI-driven exposures indefinitely, which is why the shift toward standalone AI cover is inevitable," said Lowenstein.