Calculation of probability based on population distribution

More clarity on the assumptions is needed. If the 100 people driving on the road is a metaphor for a random sample (i.e. liking peas is independent and identically distributed across the 100 people), and assuming the the 90% and 10% figures are frequentist interpretations of population probabilities, then the probability at least one driver does not like peas is one minus the probability that all like peas, which is $1-.9^{100}.$

If you think there is some "connection" between driving on the road and liking peas such that independence is not a valid assumption, then you need more information on that dependence structure to compute the probability.