Cyborg Collective [RR]

This week is all about the cyborg collective (aka, hybrid collective intelligence): how not to do it; how humans do it; and, RoboIndy.

Worse Without

Is there a colonoscopy in your near future? My wish for you is that your endoscopist is (1) using AI to improve their skills  and (2) not relying on AI to replace their skills. Unfortunately, the latter might be the trend.

A new analysis found that after endoscopists were continuously exposed to a polyp-detecting AI, their own performance on non-AI-assisted colonoscopies got significantly worse. Their “adenoma detection rate” dropped by a full 6 percentage points.

The AI turned the expert from an active participant into a passive supervisor, and their skills atrophied as a result. A tool designed to augment expertise ends up eroding it.

If your AI makes the human a simple fail-safe, you are not building an augmentation tool; you are building a de-skilling machine. This isn’t inevitable, but developers and users need to move beyond simple efficiency gains and transient metric bumps.

Technology should not only make us better when we are using it; we should be better than where we started when we turn it off again.

Creative Synchrony

What happens in our brains when we receive feedback from others? Very little good…except in one specific format.

A set of experiments on creativity and feedback found that only genuine, interpersonal feedback—a real dialogue— “improved creativity acquisition and transfer”. In fact, interpersonal feedback increases interbrain “neural synchronization" between participants, and greater synchronization leads to greater ”creativity enhancement”.

One-way feedback had no effect on creativity or neural synchronization. And needless to say, “Applying feedback positively correlated with creativity enhancement”, while “ignoring [feedback] was negatively correlated”.

The core of creative augmentation isn't just about transferring information; it's about two minds entering a shared cognitive and neural state. Right now I’m beginning my own research on synchrony in hybrid collective intelligence—between brains and machines. I predict the most creative hybrid intelligences will also show a pattern of synchrony driving creativity.

Do Androids Dream the Lost Ark?

Doctors, engineers, and Indiana Jones—the list of AI-augmented jobs grows.

A fun new article in Nature introduced a generative neural network for contextualizing ancient texts.  AI designed to help historians contextualize ancient inscriptions. “Aeneas retrieves textual and contextual parallels, leverages visual inputs, handles arbitrary-length text restoration”.

When paired with a historian, the human-AI team consistently outperformed either the human or the AI working in isolation on key tasks like restoring and attributing ancient texts.

This example of hybrid collective intelligence isn't replacing the historian; it's giving them a superpower. It handles the vast, inhuman scale of data, dynamically interacting with humans to do the irreplaceable work of interpretation, contextualization, and meaning-making. This is what a true Cyborg Collective looks like in practice.

Vivienne L'Ecuyer Ming