Against the tide: Counterintuitive findings are satisfying

Research in all its many forms is a great source of the kind of content which pushes against the narrative flow.

by Rob Corbidge

Published: 17:44, 30 November 2023

Rob Corbidge is Head of Content Intelligence at Glide Publishing Platform, applying the latest knowledge about advances and ideas in the publishing industry to our own product and helping clients get the most from their content.

Counterintuitive stories are an important piece of the content landscape. Stories that flow against the received wisdom, the current narrative, the "that's how it is".

As well as performing a vital role in providing another perspective on any given subject, or actually busting beliefs wide open, counterintuitive stories also seem to have a peculiar appeal to the human brain in a very particular way, providing a satisfying mix of intrigue and understanding.

Indeed, such was the thirst for such content at one publication I know, that if, say, dolphins could be even tenuously linked with, say, radioactive pollution or badger-baiting or arms trafficking, they would have run it as a story.

Such content doesn't need to go full flow against the tide, whatever direction that tide may be heading. It can even affirm a truth in a giant wave of contemporary confusion.

A couple of examples of this type of content came our way this week which illustrated the importance of such content, which frequently draws from interesting academic research. 

First came news from researchers at Oxford University's Internet Institute, which showed no negative impact on mental health from internet use. This is contrary to how many people feel about internet use, and in particular the use of social media platforms. 

Of course the truth there is that people are sick of some of the other people on the platforms, rather than the platforms themselves. After all, you can choose your friends but not the people you fight with in the comments.

There's a big proviso attached to such research however, and it's around access to anonymized user data. The authors point to it themselves, stating: "Research on the effects of internet technologies is stalled because the data most urgently needed are collected and held behind closed doors by technology companies and online platforms."

That said, the Oxford research itself is as thorough as would be expected and even offers gems such as "increased mobile broadband adoption predicted greater life satisfaction". You bet it does.

The dread of AI seems much of a theme at the moment, although personally I'm in agreement with the person who said something like, "The proponents of AI present themselves as wizards who can change the world, but they're more like magicians pulling a clever rabbit out of a hat." 

Consequently, it was refreshing this week to discover that humans can outperform AI systems when it comes to phishing attacks. Just. 

Why would there be any delight in research about a criminal activity that is a scourge for all businesses? Well, at present, the promise of AI comes mostly at scale, the defining word of our technological epoch. Presently, scale in this regard is an enemy of precision. One thing humans can do well is mental precision, and for those who practice such dark arts, precise manipulation.

Research from IBM pitted their splendidly named X-Force Red team of professional hackers against a generative AI model using specific prompts to produce successful phishing emails. The result was a narrow win for Team Flesh & Blood against Team Silicon, with a successful open rate of 14% versus 11%.

The team identified three factors as contributing to greater human success:  Emotional Intelligence, personalization, and finally what can be understood as comprehension of human cognition systems, manifested here by shorter and more believable email subject lines.

In other words, we know ourselves best when it comes to ripping each other off.

Knowing ourselves best ultimately points to a truth for the content industry. AI are systems of scale and summary, presently. Such systems will refine processes, and gain computational efficiency as they develop, yet at heart remain these same systems. 

AI systems have their uses, but if you really want to beguile an audience, get a human.