Sloponomics: who wins and loses in the AI-content flood?
Against all odds, the deluge might be good for creators
“Astounding triumphs”, from stopping climate change to founding space colonies, will be made possible by artificial intelligence, Sam Altman has written.
The head of OpenAI is not the only tech boss with high hopes.
AI is the “defining technology of our time”, believes Satya Nadella, Microsoft’s boss.
Mark Zuckerberg of Meta has argued that AI will enable the discovery of “things that aren’t imaginable today”.
The miracle technology’s latest feat, however, is to bring forth skateboarding kittens, wrestling grannies and a cage-fighting Albert Einstein.
Since OpenAI launched its Sora video-creation app last month, social-media feeds have filled with surreal, AI-made clips, generated by text prompt.
Meta has launched a rival called Vibes, producing similar results.
With Sora still at the top of Apple’s download chart, the internet is fast becoming clogged with uncanny AI videos that have been pejoratively labelled “slop”.
Video is just the latest flavour.
Literary slop arrived with the first chatbots: search Amazon for a new book and the results often include AI-made knockoffs with similar titles.
Image slop followed, as people shared viral pictures of Pope Francis in a puffer jacket or photos converted into the anime style of Studio Ghibli.
Audio slop has invaded playlists.
In the past year Spotify has removed 75m “spammy” tracks from its platform, saying that AI has made it easier to “push ‘slop’ into the ecosystem”.
It is unfair to call all AI-made content slop—there can be few people who have not giggled at the Balenciaga pope or given their profile picture a Ghibli makeover.
But whatever you call it, AI-produced content’s rapid infiltration of every media format poses questions about the future of creative industries.
Who wins and loses in a world of infinite content, made by countless creators?
The clearest winners are the once-frustrated artists who can bring their ideas to life.
Just as Instagram made photographers of everyone with its snazzy filters 15 years ago, AI can turn anyone into a writer, artist, composer or film-maker.
The democratisation of content-creation does not just give people a creative outlet. It provides a large minority of them with a living.
Last year YouTube paid an estimated $32bn to more than 3m creators, large and small.
Spotify paid more than $10bn in royalties to rights-holders.
AI is turbocharging ordinary people’s ability to produce, and ultimately monetise, slop of their own.
That is also good news for the platforms which share it.
Slop brings risks: Spotify warns that the listening experience is spoiled if poor-quality tracks sneak in; OpenAI admits that letting everyone generate videos free of charge is proving costlier than expected.
But helping people create slick-looking video (as well as images, sound and text) offers a chance to reverse a trend in which users of social networks have been posting less often.
What is more, platforms have a stronger hand when content is made by many creators rather than by a powerful few.
The music business has long been dominated by a handful of big record companies, which in 2017 accounted for 87% of the streams on Spotify.
Since then, following the upload of tens of millions of home-made tracks, the big labels’ share of listening has fallen to 71%.
Slop’s growing hold on attention will reduce the bargaining power of mainstream rights-holders everywhere.
Unsurprisingly, those rights-holders are the most worried.
Slop not only competes for their audience.
Many creative companies claim that it infringes their copyright.
Disney and Universal are suing Midjourney, an AI image- and video-generator, while Sony, Warner and Universal Music are suing a pair of AI audio-makers.
OpenAI breezily said that rights-holders were free to opt out of Sora; after objections from the Motion Picture Association, among others, it quickly U-turned to more of an opt-in system.
The AI industry’s general approach has been to seek forgiveness rather than permission.
It is well summed up by a video doing the rounds on Sora.
Mr Altman stands in a field full of Pokémon characters, chuckling: “I hope Nintendo doesn’t sue us.”
Top of the slops
Yet for the best-known creators, slop presents two opportunities.
First, there is little evidence that the tidal wave of user-generated content unleashed by the internet has detracted attention from the biggest stars.
Quite the reverse.
In a sea of endless content, users rely on algorithms to point them to the best stuff—and many end up in the same place.
Since 2017 the number of artists making over $1,000 a year in Spotify royalties has tripled; the number making $5m has quadrupled and that of those making $10m has risen seven-fold.
Similar patterns can be seen in the book and film industries where, despite more content sloshing around, the biggest hits are bigger than ever.
Slop’s losers will be the professional creators who are not successful enough to join the top table, and must now share the scraps with millions more rivals than before.
Second, slop offers a chance to turbocharge what creative types call discovery.
Mr Altman has touted the potential for “interactive fan fiction”, in which users can bring beloved characters to life in new contexts.
Licensing copyrighted work for sloppification could delight superfans and introduce material to new audiences.
In April, at the height of the Studio Ghibli fad, Google searches for “Ghibli” were 50 times higher than before.
Yet others have drawn the opposite lesson.
Many in the music and TV industries refused to accept that streaming was here to stay.
It took many years—and wasted billions—to cut deals with the platforms to monetise their content.
They should not make the same mistake with AI.
Where there’s slop, there’s cash.
0 comments:
Publicar un comentario