A Few Thoughts About Media In The Age Of Generative Artificial Intelligence

This technology promises to accelerate the drift initiated by the digital revolution towards a media universe characterized by a double specialization, thematic and ideological, which risks becoming a double limitation.

The digital revolution has produced a proverbial media balkanisation. By switching to bits, this market has globalized access for content consumers and advertisers, breaking down the most solid editorial monopolies. The Internet has given rise to an unlimited number of channels on every topic, lowering the barrier to entry in this field to a purely symbolic level. With social media, every Internet user has become a broadcaster, considerably reducing the competitive advantage of professionally produced content. From a sovereign sector of activity, the media have become a trivial activity for everyone.

As a result, media consumption has continued to grow and scatter: Today, Americans spend 11 hours a day reading, listening to, watching or interacting with media. More than a billion Stories are shared on Meta’s apps every day, and 694,000 hours of video are watched on YouTube every minute. Perhaps most importantly, over 39,000 TikTok accounts have at least one million subscribers, challenging the notion of the micro-influencer of yesteryear.

In the face of this popular digital onslaught, traditional media are defending themselves as best they can: Last year, 516 television series were produced in the U.S., the equivalent of one every 1.4 days. You only have to look at the strategic and ownership crisis in which Paramount Global (formerly ViacomCBS) has been embroiled to realize that the media crisis is widespread and not confined to the news media, as often believed.

Generative artificial intelligence will produce at least two effects on this media landscape:

• An even greater trivialization of news content that chatbots and other future generative human-machine interfaces will synthesize in a personalized way for each user. The news media, which complained about the use of their URL links by Facebook and Google, without accepting that it was largely advertising for their content, are going to miss what may seem like a golden age when their revenues come under even greater attack, this time in a truly predatory manner, by these applications.
• The quality of the content that every Internet user will be able to create will be on a par with the most professional productions. Tomorrow, we will no longer be able to distinguish user-generated content (UGC) from other content. From then on, the competition for attention will be more equal, and thus fiercer, than it is today.

In this context, it seems to me that niche media, which have already experienced an explosion in the Internet age (the famous “long tail” concept applied to content distribution) and Facebook groups, will thrive like never before in the age of generative artificial intelligence. This technology indeed proposes an average of human-produced content. We should not expect it, at least in its generalist applications like ChatGPT, to take any striking positions, which its statistical operation prevents, or to deep dives into subjects deemed secondary, which the endogamic use of synthetic training data would make difficult.

The triumph of political and thematic niche media, made possible by the unprecedented creative capacities available to their authors, could represent a double limitation. In political terms, the news media that will flourish in the future will be opinion-media, which will only reinforce the phenomenon observed in this respect over the last twenty years and exacerbate the prevailing ideological polarization. Thematically, we could also witness a form of radicalization, with the most incongruous leanings finding an oversized field of expression.

In both cases, a society’s ability to create common ground would be seriously threatened.