> either there is a harmful level of pesticides in our food or not
That is an impossible standard even if we had far better methods. One, "damage" is a continuum. Two, who is affected is a wide range. Three, we only test chemicals individually, and set threshold levels individually, but in reality mixing chemicals very often leads to vastly different effects. For example, when lead and mercury were tested together (LD testing) it was found toxicity increased a thousandfold compared to each one alone. But there is no way to test a meaningful number of combinations, and that would still leave out what happens in the extremely complex biochemical environment of a human body in any case.
For the foreseeable future, maybe forever even in 10,000 years of solid scientific progress (because complex stuff remains complex), will be be able to do more than do guesstimates - and they will all be on a population level, don't even try to say anything definite about an individual.
That means calling for "solid proof of damage" before anything gets banned is a sure way to stall any action almost indefinitely. I recommend a look at the history of lead (which has not yet ended at all). The most "solid proof" I can think of is a large scale and long-term human experiment, unthinkable. Even then you could raise doubts about the context/environment, methods, selected participants, etc. "Solid proof" is a lot less solid than a lot of people think, the text book examples of easy primitive physics experiments make it seem as if it's easy to find out things "for certain".