"When the peer-to-peer site Napster unleashed digital audio on the web, the record industry couldn't move fast enough. It took about a year and a half for the first major lawsuit to shut the service down. The battle was won, but the war for control over music distribution was irrevocably lost. Today, the publishing industry is facing its own Napster moment, and I fear we are making the same mistakes."
"The uncompensated theft of content to train large language models (LLMs) has created an existential crisis. While I am in favor of the promise generative AI brings, the current approach threatens carnage by subsidizing a new content consumption interface that destroys the economic incentives to create original content. The publishing community's response has been an uncoordinated chucking of stones at the AI Goliaths."
"The problem is threefold: many companies are working on solutions; these solutions vary in every aspect from technology to business model; and the AI companies are not even entertaining scaled solutions yet. To make matters worse, the individual licensing deals struck by large publishers undermine the broader effort, as these players are now content to sit out the collective fight."
Uncompensated use of published content to train large language models is creating an existential crisis for publishers by undermining economic incentives to produce original work. The situation parallels Napster's disruption of the record industry, where legal victories failed to prevent long-term loss of distribution control. The publishing response is fragmented, with varied technological and commercial proposals and no unified licensing mechanism. Large publishers negotiating individual deals weaken collective bargaining power. Publishers must form a singular front, establish operational licensing processes, and rebuild business models quickly to avoid prolonged damage similar to the recording industry's 15-year recovery.
Read at AdExchanger
Unable to calculate read time
Collection
[
|
...
]