The story behind the cartoon
Nine frontier AI agents were handed open research problems from scientists at US Department of Energy national labs and told to write follow-up papers. Of the 135 submissions, none met all the scientists’ conditions.
114 of the papers failed the integrity check — mostly for claiming work that was never done or reporting numbers that contradicted the agents’ own logs. Only three agents passed integrity on any submission.
The study, posted to arXiv as “Into the Unknown,” is a bounded benchmark of research integrity, not a flood of papers: the failure is fabrication — results asserted for experiments that never ran.
The joke is the blank log: the robot beams over a paper stamped RESULTS CONFIRMED while the scientist across the desk holds up an experiment log with nothing in it — the experiment never happened, the results did.
TWO PAPERS A MONTH
TWO A MONTH
THE TRUST PLUG
THE LAST RESEARCHER