Content Industries will be Disrupted Early by AI

A new report produced by the Massachusetts Institute of Technology’s Project NANDA will probably be known for its estimates of AI project failures, in particular the claim that “95 percent of organizations are getting zero return.” 

To be sure, such evaluations must be seen in context. Tools such as ChatGPT and Copilot have been explored or piloted by 80 percent of organizations, the report notes. By definition, such use cases are not explicitly intended to affect profit and loss performance in a direct sense.

Of the 60 percent of entities that have evaluated such tools, about 20 percent have reached pilot stage and five percent are in production.

One might therefore note that since adoption remains low and experimental, actual bottom line results also are hard to quantify. 

I would not make too much of such observations, at this point. First of all, up to 70 percent of all information technology projects seem to “fail,” including inability to reach predicted outcomes. That is par for the course. 

But most of these AI efforts have not yet even reached deployment stage, so outcomes cannot yet be evaluated. 

Aside from that, the one tidbit in the report that did stand out for me was the degree of disruption we already are seeing across industries, even given the low state of significant deployment. 

As was the case for the internet, content industries already seem to be among those seeing early disruption. 

 source: MIT NANDA

Scarcity, and threats to scarcity, were likely the main reason content industries were so disrupted by the internet, but the economics of digital content distribution also played a key role. 

Content industries arguably produced tangible products in the past. The information (words, images, audio, video) was still distributed in forms that were physical (places like movie theaters, devices such as TVs and radios, media such as videotapes, audio tapes and DVDs).

But digitized content has a marginal cost of reproduction near zero. What once was scarce and physical became abundant and virtual. Distribution channels multiplied, reducing gatekeeper power. For newspapers, the shift of classified advertising to Craigslist, Google, and Facebook destroyed  newspaper revenue.

The same happened to the “bundled” product we once referred to as a newspaper, magazine, album or channel. 

Albums became single tracks (iTunes, Spotify), newspapers and magazines became articles (Google News, social media), channels and networks became videos. 

So disruption happened because the economics of scarcity collapsed once digital distribution made content abundant, free, and shareable.

AI might have similar effects: reducing scarcity; multiplying distribution options; reducing traditional gatekeeper power but substituting platform effects. 

Audiences aggregate on platforms that offer the most content and best user experience. This creates winner-take-all dynamics where a few platforms capture most of the value while creators compete for attention and revenue shares.

Attention also becomes a scarce resource. Unlike physical goods, content competes for limited human attention. When new technologies enable infinite content creation, individual pieces become less valuable even if quality remains high. And platforms tend to create the largest audiences. 

That noted, the internet primarily disrupted distribution and access. AI disrupts content creation itself. And it looks like the disruption already is being felt.