Definitely update it... some components of IE are central to Windows XP and as a result, updating IE will make anything which uses these components more secure and possibly faster. And it doesn't take long to do.
If you’ve made up your mind not to use it then it can’t help you much, but something worth considering is that a recent study conducted independently by a third party, NSS Labs, reported that IE8 has a much higher block rate for both malware and phishing than the other leading web browsers.
I find that study highly dubious. Aside from the fact that it was commissioned and funded by Microsoft, and that the selection methodology employed to select the malicious websites is never detailed, it seems unlikely that every other browser's malware detection performance got worse between February and July. That would be unlikely even as a coincidence, and even more so considering that other browser manufacturers will also be working to improve their software.
The study claims that IE8 blocks three times more malware than Firefox, and 11 times more than Google Chrome - this seems unlikely, considering that Google maintains an extremely comprehensive list of nasty websites and is in the best position to keep that list updated! Also, Firefox and several other browsers rely on Google's index of dodgy sites - it's hard to believe that browsers using the same data feed could vary so significantly, and it's especially hard to believe that Google's own browser would perform so much worse than other browsers which simply piggyback onto Google's feed. And I find the claims that Chrome's protection 'dropped off' laughable, unless NSS are suggesting that the first batch of URLs tested happened all to be in Google's database, whilst later malware URLs coincidentally were not.
There are a lot of questions that the detailed report doesn't answer. 12,000 suspicious sites were reduced to 608 URLs which were eventually tested. The manner in which these URLs were screened is absolutely central to the legitimacy of the results, but the process used to do so is never detailed. The selection and testing process is not sufficiently transparent and I consider the report to be unduly complimentary of Microsoft and Internet Explorer, raising the possibility of bias. A previous version of this test not only included IE8 RC1 whilst using released versions of other browsers despite pre-release versions of Firefox, Chrome and Safari being available, but the resulting report also commented that 'we were impressed by the stability of IE8 RC1'. This irrelevant and unnecessarily partisan comment has to call the neutrality of the studies into question.