You are not wrong that AI is a whole new level of misinformation. But trusting the science never was a “trust any published paper” it is about trusting scientific consensus. And yeah, if there is a scientific consensus based on multiple papers and peer reviews, it is almost certainly going to be more trustworthy than your opinion/online search/intuition

Trust the science is still true, even in the face of AI, you just need to differentiate between trust scientists and trust scientific consensus.