The scientists don't even have to be shills if Monsanto controls publication of results. Pay for 15 studies from above-the-board researchers. Only publish the 6 that
didn't find harm. Later meta-studies from totally disinterested parties will also find a preponderance of evidence for "no harm," just due to selective publication practices that aren't directly visible.
At least that's how drug companies have systematically biased studies in favor of their products without needing to find amoral shill scientists. The way to prevent it is trial pre-registration, so that all trials are known about, not just the ones that turn out favorable after completion.