And does it depend on the type of field? Undergrads in biology can do the menial work, but what about undergrads in the computational sciences?
Are they more or less likely to benefit from undergraduate research in schools/fields where the undergrads tend to be especially self-motivated? (I don't know much about the ratio of self-motivated undergrads to non-self-motivated undergrads, but the professors I talked to at Brown and UChicago said that working with the undergrads at those schools was incredibly rewarding since they tended to be very self-motivated).
As you suggest in your question, this likely will vary from both field to field, and from lab to lab. In the two research labs I've worked in, one engineering and one neuroscience, undergraduates did a tremendous amount of useful work. In engineering, they would help with circuit design and fabrication, as well as doing background research and presenting their findings to the group as a whole. In the neuroscience lab, they would do cellular recordings and prepare cell cultures, as well as participate in paper writing.
In both cases, the undergraduates benefitted tremendously from the experience in a number of ways; they experienced the life of a researcher, they got to perform actual research work, they published articles and conference papers, and they received excellent letters of recommendation. The lab also benefitted, in that they had a (most of the time) highly motivated student who was interested in doing work performing research, the grad students/professors had more time either prepare other experiments or write papers, and all the benefits of simply having someone else around to bounce ideas off of. All in all, if the lab is organized enough to handle the logistics of providing the students with regular (non-busywork) tasks to perform, it's a win-win situation for everyone.