The Chi Square test is generally a conservative test. The probability of a Type I error is generally below the significance level. There are situations when it is higher but it is rarely above 0.06 when the 0.05 level is used. The Fisher Exact Test tends to be very conservative and is generally not a good idea.