Daniël Lakens

Together with my co-host @smreeteemehta we've started a new podcast: Nullius in Verba. A podcast about science - what it is, and what it could be. The introduction episode is up now! Apple: podcasts.apple.com/us/podcast/nul… Spotify: open.spotify.com/show/0tw0yVZxu… RSS: feed.podbean.com/nulliusinverba…

New blog post: Evaluating Dr. Cuddy’s Claim that the Debunking of Power Posing is a Myth. daniellakens.blogspot.com/2026/05/evalua… On an AI generated description of a non-existent study, incorrectly citing findings from studies, and the importance of scientific criticism.

We have just published an article on preregistration in parapsychology. Thanks to @lakens for great ref comments peerj.com/articles/21360/ @thePeerJ @UHertsResearch @KPUNews

Great blog post! Highly recommended

@PublicThinkOrg I am not sure how to combat this. Of course, we will soon have tools to check for fake references. But then the people who use AI will also have these tools. In the end, I guess we have little more than each other to point out problems like this.

@lakens It keeps getting worse. There's no cost to using AI unchecked, so now we can't trust the critiques to be anything better than AI slop. We need harsh consequences for fabricated citations in peer criticism. I wonder if libel laws apply here.

@MatthewBJane OK, OK, I I'll do it tomorrow! ;)

Publicly IMO. If errors are made publicly, the corrections should be public too as soon as possible so that readers are informed. Private communication can take months or years, while it continues to get cited in research, public policy, and clinical guidelines. In my view this requires an immediate public correction.

If a scientist uses AI to make strong claims in public about the scientific literature, and some references do not exist, and the generated text does not correcfly reflect the findings, should we inform them privately to fix the text or should we share this information publicly?

@Tim_T_Rex Thanks, but the claims were written online, not in a journal. And why privately? What is supposed to happen? They are required to correct the record? How do we make sure this happens to the extent I think it should? Where is the reward for the 10 hours I invested?

@lakens Privately, then the journal it is published in

@F_Bethke Only because I see no other workable solution that 1) will correct the public record 2) without inordinate time investments by me. And this is based on a lot of experience.

@lakens You seem committed to skip contacting them, so I guess go public.

@F_Bethke the people who say I should go public right away. But I appreciate you sharing your thoughts. It is important to discuss how we deal with this, but I fear your approach requires too much effort from my side to correct someone who used AI references.

@F_Bethke Yes, in Peer Review of course you have the power to require a fix, and an editor that has power to complain to. It is a different situation when others have the power to ignore you. Not surprisingly, they often do! I will end up ignoring your advice, and follow that of - >

@F_Bethke In the past 10 times you corrected mistakes of others, how often did the procedure you suggest work? And how much time did it take you?

@F_Bethke There is no journal or editor. This is a blog, by someone with close to a million followers. What does correcting look like here? How would I make sure they correct it to my liking? How much more effort should I spend on this?

@F_Bethke And if they don't, it is ok to go public? And if I start public anyway, and they do not correct it, would that confirm I made the right decision anyway?

@lakens Yes, this is what I would expect them to do.

@F_Bethke I agree with you that if I would point out these flaws, I am doing this in part to increase my own visibility and reputation as someone who sees mistakes that others fail to spot, and as someone who wants to have a healthy culture of criticism. You can call me smug. I'll own it.

@F_Bethke In this case, the claims were in a blog. If I contact them privately, what would you expect them to do? Correct the public record themselves, and made sure that their (in this case close to a million) followers see the correction?

@Acertaintom Yes, this dialogue is part of an ongoing trajectory, although it is not clear what the end result is - policies, or just more awareness?

@lakens Not sending it to a profit journal is already a sign of freedom. Be conscious that the high freedom scores in other countries may be signs of something else. Do you have a sense of when a further statement will come out?

At the Vrije Universiteit Amsterdam to join today's session on the national dialogue on academic freedom.

@Acertaintom It is a result of this letter. universiteitenvannederland.nl/en/current/new…

@lakens Is this something I could find out more about?

@BoussageonR I didn't read it, but if there is a mistake, do write a commentary. I am already spending ny criticism time this week on something else.

@lakens ...if the point estimate exceeds the set threshold, they still conclude that there is evidence of an effect EVEN THOUGH p > 0.05

The choice is not between p-values or effect sizes. They go hand in hand. The p-value is there to keep you from seeing effects where there are none. If that bar is passed, you can treat effects as non-zero, and start to reason about what they mean in practice or for a theory.