I don't think
uberrima fides is super helpful here:
a) I have a really, really, REALLY hard time understanding how hiring an army of PhDs to build highly proprietary risk models based on extremely expensive (or not-even-for-sale) datasets does not run a foul of uberrima fides. The doctrine only makes sense if it forbids both parties, not just the insured, from concealing information. If I were on a jury and the entire case hung on a reasonable man's interpretation of uberrima fides, I would have a hard time ever finding in favor of a modern insurer who's unwilling to share their models and data with the world. I mean, I might agree that the insured hid information. But I'd be nearly 100% confident that the insurer hid information.
Maybe in 1766 this principle made sense. It aint 1766 anymore.
b) In this case specifically, it would be quite hard to convince me that using publicly available information (e.g., weather reports and history of on time / delayed / cancelled flights) runs a foul of uberrima fides. You can't honestly expect me to believe the insurance company didn't have access to that information. It's equally hard for me to believe that an army of PhD actuaries didn't think to use that information to build their pricing models.
c) In the case of travel insurance, the doctrine has a really big bright-line problem. When I buy any ticket into or out of Boston during the winter months, I tend to buy travel insurance. I never buy travel insurance for flights into or out of Boston in the summer. That is clearly not fraud (or, if it is, that insurance product needs to be regulated out of existence). Now, what if I choose particular weeks? days? Where's the bright line?
d) Even supposing some obligation to share information, how in God's name am I supposed to inform the insurer that I estimate a flight will be almost surely be cancelled/delayed? Call up the 1-800 number? Because I know exactly how that would go: you'd get a first-line support person who's incentivized to sell policies. If I were running a similar scheme, I would definitely record myself calling up the insurance provider and point-blank stating "I think this flight will be cancelled, should I buy your insurance?" I guarantee the answer from the T1 support folks would be "yes, that's what it's for!". It seems totally unreasonable to assume bad faith of anyone who buys an insurance product with large positive expected value, especially when the insurance company is consistently making 10 figures in net income and is unwilling to engage in individual dialogs about "who knows what and when". If I can't get on the phone and talk to the actuary who built the pricing model to figure out if X is shared information, and if there's no list of check-boxes for me to look through in order to determine the same, and if the only company representative I can get access to is probably going to explicitly tell me to buy the policy, then... what the hell?
All of these things taken together: if I were on a jury and uberrima fides was the only reason to find in favor of the insurer, I would certainly do everything I could to sway my fellow jurors toward favoring the insured.
It seems like the right solution here is either a) more sophisticated pricing models, or b) some very clear and well-dilineated bounds on what information is/is not allowed to be used when purchasing travel insurance.
TL;DR: uberrima fides as a fuzzy legal doctrine seems... just... utterly impossible to apply to consumers. The relationship is fundamentally asymmetric, so the insurer should have a burden to explicitly ask about any relevant information (as is done for, e.g., health and life products). As a fuzzy doctrine for deciding on a case-by-case basis, this doctrine seems more relevant when the insurer and insured are more symmetric (e.g., insurance products aimed at large corporations, reinsurance, etc.).
(Of course identity fraud is a problem and this happened in China which doesn't use the same legal system, so this is all just a hypothetical conversation.)