Doesn't need to be AI to be garbage per se, popular science should always be read with some skepticism as it's often a translation (pop sci) of a translation (news publication) of a paper (the actual source).
Example, pop sci will show a rendering of a lush green planet with a headline like "EXTRATERRESTRIAL LIFE FOUND!11", the news release will be something like "Potentially life-supporting planet discovered by xyz", and the actual paper will be something like "measuring device foobar123 noticed a 0.2% decrease in luminance of star aybabtu-1337 at a period of .6 frotz/picoyear indicating this planet probably isn't cooked or frozen but what do we know lol"
Example, pop sci will show a rendering of a lush green planet with a headline like "EXTRATERRESTRIAL LIFE FOUND!11", the news release will be something like "Potentially life-supporting planet discovered by xyz", and the actual paper will be something like "measuring device foobar123 noticed a 0.2% decrease in luminance of star aybabtu-1337 at a period of .6 frotz/picoyear indicating this planet probably isn't cooked or frozen but what do we know lol"