Data from the MIT AI Risk Repository shows the frequency of AI-related harms reported globally across 1700 risk types.
The number of reported AI incidents has increased significantly in recent years. Risks associated with malicious actors are the largest incident category. System failures, misinformation, and discrimination are also top concerns. In turn, litigation related to AI-driven harms is becoming increasingly common.
Traditional insurance policies don’t often provide comprehensive protection against AI-related risks. Certain cyber, technology errors and omissions, and general liability policies may cover some AI exposures, but they frequently contain exclusions that leave organizations vulnerable.
Implications for brokers and their clients:
- Investigate dedicated AI insurance that explicitly address emerging exposures, including autonomous decision making, and cyber and IP risks.
- Review existing cyber, tech E&O, professional liability, and general liability policies to identify potential AI-related exclusions.
- Consider strengthening D&O insurance in case of AI-related governance failures, regulatory investigations, and shareholder claims.
Source: EconoFact (May 6, 2026). Can Companies Insure Against AI’s Growing Risks?
Emerging insurance industries mentioned: Artificial Intelligence Insurance.
Lines of business mentioned: Cyber Liability Insurance, Tech E&O Insurance, Directors and Officers Liability Insurance, Errors and Omissions Insurance, General Liability Insurance.
AI-Driven Fraud Threat Escalates in Germany
The growth of AI is increasing the risk of white-collar crime. A KPMG study of German companies found that:
- 48% of firms rate the risk of AI to their own organization as high or very high.
- 73% consider the risk of AI-related data theft and misuse as high or very high.
- 71% consider the risk of AI-based fraud as high or very high and 58% expect it to increase over the next two years.
- 65% have reported incidents involving fraud and embezzlement.
- 51% were affected by the disclosure of trade secrets.
31% of companies cited losses of less than €10,000 but for 18%, that figure reached from €90,000 to €100,000.
On the plus side, the detection of white-collar crime is becoming more systematic. Accidental discovery now accounts for only 5% of cases, down from 50% in 2023. Instead, 30% of incidents are being identified through annual audits and 31% through tips from regulators and law enforcement.
Implications for brokers and their clients:
- Review first and third-party cyber liability insurance to ensure AI-enabled cyberattacks are covered.
- Review IP protections to ensure coverage is in place for AI-enabled theft or misappropriation.
- Review crime insurance policies to determine whether they provide sufficient protection against AI-enabled fraud, including social engineering attacks and deepfake impersonation.
Source: KPMG (June 2, 2026). White-collar crime in Germany: Companies see rising case numbers due to artificial intelligence.
Lines of business mentioned: Cyber Liability Insurance, Crime Insurance.
Minnesota Unifies Cannabis Tracking as New Medical Rules Take Effect
Minnesota has updated its cannabis regulatory framework through a newly enacted Omnibus Cannabis Bill. Signed last week by Governor Tim Walz, it affects both hemp and cannabis firms. Some changes are effective immediately and others will come into force on January 1, 2027.
Firms will no longer have to maintain separate medical and adult-use supply chains in the Cannabis Tracking System. License holders with medical endorsement can now manage cultivation, production, and inventory through a single system, with unified tracking in Metrc.
The current medical cannabis combination business license will be replaced by a “macrobusiness” license. This will permit:
- Up to 28,000 square feet of indoor canopy. This is a vast reduction from what the previous license offered, but compliant operators will be able to request canopy increases over time.
- Up to eight retail locations (location requirements will be tied to areas of high need).
- A cap of eight licenses throughout the state until January 1, 2030.
Current medical licenses must be converted by January 1, 2027.
Medical-endorsed companies will have to employ a licensed pharmacist or medical cannabis consultant. They’ll also have to offer 24-hour fulfilment options and other priority service measures and will be mandated to stock products of high medical need.
Reporting obligations will be streamlined. Operators will have to submit an annual market analysis and consolidated legislative report by January 15.
Implications for brokers and their clients:
- Consider strengthening errors and omissions insurance to provide adequate protection against professional liability risks, including dispensing errors and product-related advice.
- Consider business interruption insurance to help mitigate losses arising from operational disruptions or temporary suspensions associated with licensing changes.
- Investigate dedicated cannabis insurance solutions that are specifically designed to address the sector’s unique risk profile.
Source: Harris Sliwoski (June 3, 2026). Minnesota’s New Cannabis and Hemp Laws.
Emerging insurance industries mentioned: Cannabis Insurance.
Lines of business mentioned: Errors and Omissions Insurance, Business Interruption Insurance.
Private Key Breach Triggers $30 Million in Crypto Losses
Humanity Protocol has experienced a security breach involving an estimated loss of more than $30 million, according to on-chain analytics from Lookonchain. Following the exploit, its H token declined by 90%.
The stolen funds were dispersed across several wallet addresses believed to be controlled by the attacker. Reports indicate that private keys linked to a member of the Humanity Foundation were compromised.
The attacker continued converting the tokens into Ethereum, which has prevented any significant price stabilization.
Implications for brokers and their clients:
- Consider digital asset crime insurance so that losses arising from private key compromise, wallet theft, or unauthorized asset transfers are covered.
- Consider smart contract failure insurance to protect against coding errors that lead to losses.
- Consider dedicated digital asset and web3 insurance that safeguards firms from the sector’s exposures.
Source: CoinPedia (June 9, 2026). Humanity Protocol Hit By $30 Million Exploit As H Token Crashes 90%.
Emerging insurance industries mentioned: Digital Asset and Web3 Insurance.
Lines of business mentioned: Digital Asset Crime Insurance, Smart Contract Failure Insurance.
New Zealand Tightens Grip on Online Gambling with New Harm-Reduction Rules
New Zealand is moving ahead with the implementation of its Online Casino Gambling Act 2026. A set of regulations were published last week which will come into effect on July 3. The measures impose new obligations on operators, including responsible gambling controls, tighter advertising standards, and a levy on gambling profits.
Under the new rules, players must be given tools to manage their gambling activity including configurable limits on spending, deposits, and time spent on a platform. Operators will also have to introduce mandatory ‘break-in-play’ features, process self-exclusion requests within 24 hours, and support customers showing signs of gambling-related harm, which may include restricting their access.
In terms of advertising, the following will be banned:
- Ads on front pages of print publications and on public transport.
- Ads that encourage impulsive gambling behavior.
- Ads appealing to minors.
- Personalized ads that are designed to increase spending.
- Broadcast ads are shown within 30 minutes on either side of live broadcasts.
- Inducement advertising (unless under strict control).
- Sponsorships, endorsements, and affiliate marketing.
In addition, operators will face stricter ID verification requirements and new payment controls. For example, customers will only be allowed to connect one payment method to each platform.
Implications for brokers and their clients:
- Consider cyber insurance to help protect against losses arising from enhanced identity verification requirements.
- Consider D&O insurance so that executives are covered for regulatory investigations and governance-related claims stemming from the new requirements.
- Consider specialized gambling insurance that addresses exposures including player liability, payment fraud, and regulatory risk.
Source: iGamingBusiness (June 8, 2026). New Zealand introduces online gambling regulations ahead of licensing process.
Emerging insurance industries mentioned: Gambling Insurance.
Lines of business mentioned: Cyber Liability Insurance, Directors and Officers Liability Insurance.
Illinois Sets New AI Transparency Rules While Liability Debate Continues
Illinois recently became a key testing ground for AI safety regulation. Two competing bills offered different approaches, but SB315 has now been passed.
SB3444 (the Artificial Intelligence Safety Act) would have provided a safe harbor for developers of the largest frontier AI models. Developers would not have been liable for catastrophic harm if they complied with certain safety and transparency requirements and didn’t act intentionally or recklessly.
The bill defined ‘critical harm’ very narrowly: incidents involving the death or serious injury of at least 100 people, or at least $1 billion in property damage caused or enabled by the model through chemical, biological, radiological, or nuclear threats, or autonomous criminal conduct.
It would have only applied to models that have a compute cost of over $100 million or those trained with over 10²⁶ floating-point operations. Any firm building on top of those models would have fallen outside the act.
SB315 takes a different approach. It focuses on transparency, auditing, and disclosure requirements for catastrophic risk, but it doesn’t create civil liability.
Both bills highlight a discrepancy between reality and legislation. OpenAI is already defending an array of wrongful death lawsuits, but these harms would have fallen outside the scope of SB3444 because they involve individual victims rather than mass-casualty events. SB315 avoids granting safe harbor to developers, but it still doesn’t create new avenues for victims to seek compensation.
Implications for brokers and their clients:
- Review general liability and product liability coverage to ensure adequate protection against third-party bodily injury, property damage, or other harm allegedly caused by AI systems.
- For firms developing applications on top of frontier AI models, consider evaluating third-party liability protections related to the deployment, customization, or use of those systems.
- Consider reviewing regulatory defense cost coverage to help manage the costs of responding to AI-related investigations and enforcement actions.
Source: Forbes (May 27, 2026). Where Does ‘Catastrophic’ Start When It Comes To AI Liability?
Lines of business mentioned: General Liability Insurance, Product Liability Insurance.
