I think the answer has to be indeterminate, as must many questions debating ethics without specifying a code of ethics to use. There's going to be a mix of positive and negatives, and its always up to the individual to decide which is which.
I often argue that science is an exceedingly violent arena for ideas. Science sells a sort of king-of-the-hill approach, where less than perfect ideas are sacked, providing fuel to propel more perfect ideas to the top of the hill. We see this in the version of the scientific method which is taught in schools. When you have hypothesis, you don't directly test it. Instead, you identify a null hypothesis and prove that that null hypothesis is wrong. You then suggest your alternate hypothesis, which fits the data, is a better one. Typically the null hypothesis is a currently accepted theory, so disproving it kills off that theory, and your theory propels itself forward over the dead body of the last theory.
Science often makes ontological claims which are at odds with its empirical underpinnings. Rarely will you hear a scientist carefully state, "the models of motion which most accurately align with empirical evidence suggest that there is a force called gravity that pulls things down. Instead you will hear "what goes up must come down," which is an ontological claim that makes the assumption that what has occurred in the past will occur in the future.
How do they get away with this? Well, most often they're right. We've built our technological society around the assumption that the science is most often right. To defend this claim, we turn to the king of the hill model that science proscribes. We take solace in the knowledge that the currently accepted theory is not just any theory. It's a survivor that is being constantly challenged from all angles, and so it is at least as good as the combined scientific community can possibly provide.
Why do I say all this? Because science has a code of ethics to support this. In the scientific community, you are supposed to publish your responses, even if you don't like the outcome. You are supposed to publish, even if it might harm a handful of people. This is essential in the medical community, where an experimental procedure might save 1 or 2 lives, but not knowing that the procedure is flawed may kill thousands once the procedure is approved.
So at the very least, if one is doing a utilitarian analysis, one must be able to weigh the deterioration of the ethics of science against the ethics of hurting people (such as those eating too many bananas for their own good).
These are the ethical decisions to be made with science. If you go to other disciplines, the rules of ethics may be different. Traditional Chinese Medicine (TCM), for example, has a very different code of ethics which supports the way they approach healing. I do not practice myself, so take my opinion with a grain of salt, but with them, the health and wellbeing of their patient is their foremost concern. Having the "true" answer is more of a convenience to them, making it easier to care for the health and wellbeing of their patient. If a placebo effect makes their patient better (in a holistic sense), then that is an acceptable approach to them. As a result, you see that the things built on TCM have a very different flavor than the things which are built on Western science, which is very dependent on knowledge. (If you look at what goes into a surgery with general anesthesia, you appreciate just how absurdly dependent on our knowledge we are!)