Discussion about this post

User's avatar
Woolery's avatar

Effective misinformation will adopt whatever tone/communication standard that its targeted audience is accustomed to associating with valid information. Your example of “enhanced interrogation” reflects this. In this sense the proposed inoculation technique wouldn’t be effective against purposefully, well crafted disinformation, as you mention.

I think it’s also worth pointing out that since van der Linden’s book, LLMs have made it much, much easier to disguise any of the telltale “fingerprints” of misinformation that are often just symptoms of poor communication technique combined with outsized enthusiasm on an issue.

All that said, a headline that reads “Wicked, corrupt fascist leader conspires to eat beloved pets of true patriots,” should throw up some misinformation red flags, and though language this obviously crafted as propaganda can make unstable individuals act in dangerous ways, I don’t think it’s the primary problem.

Expand full comment
Matt Grawitch's avatar

Dan, at the expense of joining the "wrong side" of this debate, I tend to agree with the vast majority of your post here (I, too, am suspicious of the inoculation claims, so I was obviously biased in that direction to begin with). Part of the problem I have is that a lot of the inoculation research is attempting to help people develop heuristic rules to spot (and subsequently reject) misinformation without actually establishing that these rules are helpful for differentiating anything that isn't at one end of the extreme or the other. In other words, show me evidence that these heuristic rules are helpful for differentiating true from misinformation as we approach more of the gray area end of the spectrum (though spectrum is not really all that accurate).

But another one of my big problems is that "misinformation" has become this umbrella category that has lot any sense of meaning (if it ever had any to begin with). The simplistic categorization of content into misinformation vs. true information is overly simplistic. What degree of inaccuracy is required for something to be categorized as misinformation (how much unsupported premises/claims, faulty logic, cherry-picking evidence is required)? Is one bad part of an argument sufficient? Does it require to come from sites like InfoWars (where the misinformation may be blatant and recurring)?

Anyway, I won't write a lengthy essay-length response here, but from a strictly decision-making perspective, I would tend to agree that a lot of your arguments line up well.

Expand full comment
40 more comments...

No posts