We would likely not have a large enough risk pool in this instance. A large risk pool is one of the conditions for a risk to be insurable hence the product would likely not be offered, in theory.
However, there are instances where a risk with a small risk pool is nonetheless insured. If such a product is being designed and there is an indication of low demand then it would be priced higher to ensure that the costs can be covered (unless the insurer was happy to make a loss on this product, such as with loss leaders). Products are priced so that the expected premium income is greater than the expected payouts. The primary reasons the premium would be higher is that the fixed costs would now need to be spread among fewer policies, and, you'd likely also add some risk margin for the additional risk involved with having a small risk pool, which also increases premiums
If the insurance product is already on the market but few people are buying it then the insurer could potentially keep the premiums low while they try to get more policyholders (if they have the capital reserves to sustain this). They could also lower premiums, but the premiums would still need to stay above the expected payouts for the policyholder otherwise they'd likely be worsening their position. Increasing the premium is an option too but the decision would be dependent on the level of elasticity of the product. For more inelastic products, the increase in price will be more effective in increasing premium income (without increasing the liability to the insurer). While decreasing premiums may be more effective for elastic products.
The insurer may also choose to cancel the product if they believe they are not in a position where they can provide the cover. They may then return some of the premiums to policyholders who had bought cover, and then alleviate themselves of their liabilities.
Okay, that feels like plenty to read. Hope it helps.