Save it and forget it, for better or worse

Hats off to UCL team of researchers, Chiara Scarampi, Chhavi Sachdeva, Pei-Chun Tsai, and Sam J. Gilbert for their perceptive study providing a neural explanation of "digital amnesia" using fMRI technology and multivariate pattern analysis.

Read the study here

The team put 36 volunteers inside an fMRI scanner, a non-invasive medical imaging device measuring and mapping brain activity by detecting changes in blood flow and oxygen levels.

The volunteers viewed sets of pictures, faces, and scenes, following which they got one of three instructions: a) Remember this, a test could follow. b) Forget this, no test. c) Offload it: if a test follows, answers will be provided.

The researchers then compared the brain activity triggered by each instruction, and tracked the neural trace of the pictures to see the effect on memory following the instruction.

The brain activity for (c) looked far more like "forget" than "remember". For (a), the signature of the pictures stayed readable in the brain right up to the test. For (b) it faded until it was statistically absent, akin to deliberately forgotten material.

This is reportedly the first study to decode the brain mechanism behind offloading, which points to deliberate forgetting. But the researchers read this as design, not decay. A brain that clears what's safely stored elsewhere is making room, and people who offloaded one set of pictures remembered the next set better.

The study used automatic reminders, not AI, but what happens when we hand a half-formed question to a chatbot before wrestling with it? Forgetting a saved phone number is fine; can't say the same about shunning a half-formed question. Can we? Especially given how many folks are moving fast on a self-defeating pathway: from cognitive offloading to cognitive surrender (not merely decline.)

The research is a work in progress and we keenly look forward to the insights to follow, especially focused on offloading to AI rather than to static reminders.

Meanwhile, here are a few questions, half formed in our minds, which don't let us shun them (for good, we believe):

• How closely do automatic reminders mirror real-world offloading?
• What is the strongest evidence that offloading is psychologically distinct, and not just a relabeled forget instruction?
• How many people did the screening exclude, and might people who resist or ignore reminders show a different pattern?
• What experiment could help know whether offloaded memories are erased, suppressed, or merely demoted?
• Does the activation pattern give a hunch about what people were actually doing in their heads, and could a study catch those strategies in the act?
• Given how unreliable AI assistants can be, how does the brain offload to a tool it cannot fully trust? How does one study what happens to the human brain following years of heavy device dependency and AI use? Do they lose the urge or motivation to use the freed-up space for something better?