AI-generated image from Bridging the gap: explainable ai for autism diagnosis and parental support with TabPFNMix and SHAP, by Shimei Jiang, published 2025-11-19 in Scientific Reports, a Nature journal.
There is reason to believe that much of the above article was generated by LLM AI. How did this get through the peer-review process at a reputable scientific journal? I don't think we will ever hear the details, but the error wasn't discovered until readers noticed the absurd image after the paper was published.
The Russian proverb, "Trust, but verify", is good advice but very incomplete. Excessive trust may become obvious after the fact. But none of us have time for complete verification of everything we see.
This isn't a new problem. The public Internet has always been full of incorrect, incomplete, and/or biased information. Ditto for broadcast media since the beginning of radio, newspapers since the beginning of printing, and word-of-mouth since the beginning of words. AI just adds a more efficient way to produce believable nonsense.
Here are a few resources for truth-seekers:
WolframAlpha has a mission: "to collect and curate all objective data; implement every known model, method and algorithm; and make it possible to compute whatever can be computed about anything." If you want to know the volume of 200 grams of rice, or how to integrate ex, look here.
The Online Etymology Dictionary is "a map of the wheel-ruts of modern English. Etymologies are not definitions; they are explanations of what words meant and how they sounded 600 or 2,000 years ago." But this can tell you a lot. Next time you hear somebody claiming that "handicap" comes from "cap in hand", you can discover the truth here.
Who fact-checks the fact-checkers? Well, you do. But here are a few with good track records:
Snopes is the best-known fact-checking site. They aren't always perfect. But their articles usually include links to their sources, so you can do further research and come to your own conclusions.
Lead Stories is a rapid-response fact-checker: "we specifically hunt for trending stories, images, videos and posts that contain false information in order to fact check them as quickly as possible." They're a good place for verifying viral posts on social networks.
Science Feedback has in-depth articles about scientific misinformation.
Retraction Watch is a blog and a searchable database that covers scientific journal articles which have been retracted.
Internet Archive is always a good place to start if you're looking for old stuff.
The Data Rescue Project is a central node in the large network of independent efforts to save information that governments want to erase. Much of this information was tossed frantically into the lifeboats, as erasures ramped up in early 2025, and this site provides a search tool for locating it.