You are likely already aware that the IAT has potential to reveal troubling aspects of 'human nurture.' It therefore poses significant possibilities for misuse. If you are considering use of the IAT in research, this source of concern will of course be taken into account in developing your research plans, in accordance with safeguards present in the institutional setting of your research.
Although the IAT was developed for research use, it has clear potential for application outside the research laboratory. The following possible misuses might arise when operating outside the laboratory (and therefore operating beyond the scope of safeguards present at research institutions). First, people may use the IAT to make decisions about themselves: what should I buy, where should I go to school, etc. These seem, at least on the surface, to be acceptable (I may use any method I like, including looking at tea leaves, to decide that I want to work at Firm X, Y or Z rather than at Firm A, B or C). Second, people may use it to make decisions about others: for example, one use might be to ensure that people showing a certain degree of automatic racial preference cannot participate in decision-making in racially sensitive situations. Third, people may use it to investigate others' likes and dislikes, causing them harm in the process. Project Implicit, along with the investigators who have been involved in developing the IAT, urge careful consideration of costs associated with these possible misuses in contemplating any applications of the IAT.