Shahan Ali Memon

Failures of retraction are common in science. Why do these failures occur? And, relatedly, what makes findings harder or easier to retract? We use data from Microsoft Academic Graph, Retraction Watch, and Altmetric---including retracted papers, citation records, and Altmetric scores and mentions---to test recently proposed answers to these questions. LaCroix et al. (2021) employ simple network models to argue that the social spread of scientific information helps explain failures of retraction. One prediction of their models is that widely known or well established results, surprisingly, should be easier to retract, since their retraction is more relevant to more scientists. Our results support this conclusion. We find that highly cited papers show more significant reductions in citation after retraction and garner more attention to their retractions as they occur.

Recent proposals urge AI labs to prepare for “AI welfare” under uncertainty about whether AI systems have morally relevant inner states. We do not argue for or against the possibility of AI welfare. Instead, we argue that current AI welfare assessment fails for two linked structural reasons absent from other evaluation targets. First, AI welfare indicators are co-engineered with the systems they evaluate: ordinary development decisions that shape model behavior can also manufacture or suppress welfare evidence. Second, AI welfare lacks external validation: no deployment failure or independent test can reveal whether a welfare metric tracks anything real about the system. Together, these problems yield our central claim: For current systems, AI welfare is bullshit in Frankfurt’s sense, as its measurement regime is structurally disconnected from truthtracking. AI welfare should therefore not be institutionalized as a binding gate for oversight, release, or accountability; restrictions on AI systems should instead be justified by externally verifiable harms.