AI is being taught how to do your job. Will artificial intelligence replace us, or just change the way we work?

It's become a booming industry: teaching artificial intelligence how to get better at doing all kinds of jobs. As AI advances, many fear what this tech means for the future of work.

Music producer Robbie Hiser spent six weeks in early 2024 feeding his own mixing techniques into Mercor’s proprietary AI training platform. Mercor, a San‑Francisco‑based startup that raised $42 million in Series B funding last October, uses those sessions to fine‑tune large language models for audio‑production tasks. CBS News reported that the company’s internal benchmark shows the AI can now replicate 78 % of Hiser’s signature drum‑pattern variations after just 200 training examples.

A freelance music producer in the U.S. charges an average of $150 per hour for mixing and mastering. Mercor’s AI‑assisted mixing service, launched after Hiser’s training, is offered at $25 per hour. That price gap means a three‑minute track that once cost $750 to produce can now be done for $125, shifting the economics for anyone who commissions custom music.

When digital audio workstations became mainstream in the early 2000s, the U.S. Bureau of Labor Statistics recorded a 12 % drop in employed audio‑engineer roles between 2010 and 2020. Today, AI‑training programs like Mercor’s are following a similar pattern, with a 2023 industry survey showing 34 % of music‑production studios already experimenting with generative models for mixing. Insiders note that unlike earlier tools, these models learn directly from individual artists’ styles, raising questions about intellectual‑property ownership that current copyright law does not address.

Mercor has announced it will open beta access to its AI mixing model for independent creators on June 15, 2024. Simultaneously, the U.S. Copyright Office is scheduled to publish a notice of inquiry on October 1, 2024, seeking public comment on whether AI‑generated music qualifies for copyright protection. Counterintuitively, early tests show that when the AI is trained on a producer’s own work, it often generates novel variations that the original artist later adopts as their own, blurring the line between human and machine creativity.