OpenAI has missed its own 2025 deadline on a tool it said would allow photographers to exclude their work from the company’s training data.
Back in May, OpenAI revealed it was working on a tool called Media Manager that would “identify copyrighted text, images, audio, and video.” The proposed program was intended to quell some of the company’s criticisms and potentially shield it from its many copyright disputes.
But news about Media Manager is non-existent with one former OpenAI employee telling TechCrunch they don’t believe it was a priority for the company. “To be honest, I don’t remember anyone working on it,” the anonymous employee says.
After DALL-E 3 was announced, OpenAI’s image generator, the company announced that photographers could opt out of AI training data. “We understand that some content owners may not want their publicly available works used to help teach our models,” OpenAI said at the time.
But photographers had to submit each piece of work they wanted excluding along with detailed descriptions, a cumbersome process.
Similarly, Media Manager would also put the onus on creators to remove their work, something that Ed Newton-Rex, the founder of Fairly Trade, says is unfair.
“Most creators will never even hear about it, let alone use it,” he tells TechCrunch. “But it will nevertheless be used to defend the mass exploitation of creative work against creators’ wishes.”
AI and training data is an extremely thorny issue with many photographers and content creators of all stripes feeling aggrieved that their work is being used to build products without ever giving permission or being remunerated.
Media Manager was initially pitched as a way to sort out these problems and offer photographers an easy way to opt-out. But despite setting itself a deadline of “by 2025”, OpenAI has stopped mentioning Media Manager altogether. The last update was in August when a spokesperson told TechCrunch that the tool was “still in development.”
An IP attorney tells TechCrunch that Media Manager is an ambitious project; noting that large platforms like YouTube and TikTok struggle with content ID systems.
“Ensuring compliance with legally required creator protections and potential compensation requirements under consideration presents challenges,” says Andrian Cyhan from Stubbs Alderton & Markiles. “Especially given the rapidly evolving and potentially divergent legal landscape across national and local jurisdictions.”
Image credits: Header photo licensed via Depositphotos.
