The concept of science osmosis is a cool and interesting mechanic but its currently implementation unfairly punishes high influence with no regards for whether the civilizations involved have any differences in science output.

The current system grants a city bonus science based on its own science output when it is being influenced by another civilization. It does not take into account the other civilization's science output and pays no regards to whether the civilization being influenced already has more technologies unlocked than their influencer civilization. This leads to situations where a civilization with a higher science output can double the science output of a city through absolutely no effort and becomes absolutely broken in situations where a player focuses on science output (or utilizes the science affinity effect to convert an entire city's output to science).

For example, my friend an I found ourselves in a situation where he was playing as the French and I had achieved cultural domination of the entire map. He had 65 technologies researched to my 60, and there were no technologies that I had researched that he hadn't. By using Collective Minds on his main city, he reached an output of 20,000 science per turn in that one city. In addition to this, he was receiving osmosis events every turn that were then granting him an additional 20,000 science per turn on that city. Needless to say the game ended rather quickly because he reached a science end condition far quicker than I could react.

Thematically this made no sense whatsoever as his civilization was already far more advanced than my own. Gameplay-wise, this meant that I could not ever compete with his output even by choosing the Swedes (I did not choose the Turks because we have a list of civilizations we avoid) because the game was actively rewarding him for a decision I made.

The sharing of science between cultures and the competitive nature of trying to adapt your technologies to those of your neighbours makes sense and is a cool mechanism that should help players who are near more advances ones, but this is not currently how the mechanic operates. science osmosis should not be based on influence (or at least not exclusively so), because this discourages players from investing in a resources that is already extremely difficult to accumulate. It makes no sense thematically that a more advanced culture should continue to be rewarded while a more culturally dominant player falls behind, and it also makes no sense that the bonus should be based on the influenced player's city output - particularly in cases where their output exceeds that of the other player's entire civilization. The sharing of knowledge and ideas isn't one-sided, and the current mechanic is unintuitive, unrealistic, and punishes players for investing in a resource that isn't even directly related to science (influence).

I agree that science osmosis needs more clarification and prevent a more advanced culture from receiving boosts from less advanced cultures around it. I am way ahead in the tech tree than my neighbour edo japanese but for some reason i receive science osmosis from them?

Also I can totally see civics that prevents science osmosis (patenting inventions) or even an espionage system that targets the stealing of advance technology to add twists into the current system too.

Yeah the current system is broken by two things:

A) It is exactly as you described. I had a game were i was far more advanced by everything, just didn't research 1-2 early techs. Got this event and so much science

B) There is another event which lets you chose to get some advance in a tech you do not have and just free science points. The free science points are always more and lets you advance further in the direction you want.

Civ IV and V had a much better solution. Based on the known civilizations which already got the tech, you get a percentage science bonus when researching that tech (and that tech only, any overflow to another tech get reduced first, so no exploits possible).

The free option seems to be 25% of your current research.

The current system is not realist and it is exploitable in the early beginning if you volontary found your first city as close as possible with a leader beneficing +5 influence by city by turn trait.