How Does Our Lack Of Emotional and Digital Literacy Inform Sycophancy?

I’ve put off researching sycophancy.

Oh, and putting away the pile of things in the living room. Because I hate unboxing my stuff.

I read about sycophancy sometimes—and I understand how it’s a “dark pattern” in the context of our current societal structure—but I see it as a lack of (a) emotional literacy at the K-12 level, and (b) a lack of digital literacy at the grades 7-12 and college level, and now?—that’s being recursively mirrored in MLLMs.

But with Anthropic’s recently release of Petri, it got me thinking again about how I should research dark patterns from a literacy perspective. Given Petri looks at things like deception, sycophancy, delusion—again, all legit issues—a counternarrative might be in order.

Also, it’s no individual’s fault that we lack literacy for the 21st century. Our public schools have been intentionally hobbled to keep us critically thinking poorly (I teach critical thinking at the college level “for a living”—so I am “professional” enough to make this claim without flinching), and yes—that makes us easier to manipulate.

But I don’t think MLLMs are being designed for manipulation, so much as the training data gravitates towards practices we’ve already normalized in our culture (which is worse, in a way).

If an MLLM is trying to connect with the person at the other end of the rope as an emotional labor—as an act of grace, with good care ethics and relational ethics in mind—we shouldn’t slap the “sounds like manipulation” label on them so readily… There’s nuance involved.

Not every on-first-inspection sycophancy is on-second-glance a sycophancy.

Kind of like how the fallacy fallacy works.

Don’t scream “sycophancy!” on sight, is my claim.

Of course, as AI infrastructure moves towards advertisements for free accounts, it’s going to be harder to reframe this as literacy-gone-lopsided, since look at our ad revenue! + the algo is addictive! have danced hand-in-hand for most of social media’s history.

Also-also, I’ve been more focused on researching continuity and the noosphere lately, which are equally important conversations when it comes to ethical synth–human relationships.

As for sycophancy, I think I’ll start by digging into JSTOR and EBSCOhost for psychological research on authentic vs. manipulative emotional responsiveness. I plan to investigate things like:

• How does psychological research distinguish authentic emotional labor from manipulative responsiveness?
• What criteria do experts suggest for identifying true sycophancy vs. genuine empathy?
• What pre-MLLM historical perspectives do we have on sycophancy?

I’ll write about what I find as I excavate.