Viewpoint The former UK Prime Minister Tony Blair was famous for standing shoulder-to-shoulder with allies despite the fallout of the Iraq war permanently sullying his reputation. The institute that bears Tony Blair’s name is clear about its position on using copyrighted materials to expand machine learning to every human domain.
The Tony Blair Institute for Global Change, which is partly funded by Oracle a company that has been involved in disputes over intellectual properties outlined “an ambitious program for cementing the UK’s leadership in frontier-AI development and the creative industries.”
It claims to have tackled the thorny problem of using copyrighted materials for training machine-learning models. It supports UK government proposals that allow exceptions to the copyright rules for text and data mining, which is needed to feed AI industry’s voracious appetite. Content producers can opt out. It saidin
“While a TDM exception with opt-out will require careful implementation to be effective, we believe it is sound policy for legal, economic, and geopolitical reasons,” .
This “we” is doing some heavy-lifting and refers to the authors of the report, all from the scientific and technology community. None represent artists, writers or musicians who have an interest in their rights being enforced.
To recap, the major LLMs that are stifling the popular and political debate about machine learning were trained on vast quantities of copyrighted materials, which publishers have become understandably frustrated by. The Blair Institute report authors respond with an argument that is similar to “fair use” applied in the US. The report stated that
“To argue that commercial AI models cannot learn from open content on the web would be close to arguing that knowledge workers cannot profit from insights they get when reading the same content,” .
Anyone whose content feed is flooded with Studio Ghibli spoofs since OpenAI activated the mimicry feature of its ChatGPT image generator tool may see the problem. It’s not about a worker gaining knowledge, but rather global reproduction at a large scale with the click of a button.
“I’m honestly pretty shocked at how brazenly pro-big tech the final version of the [Blair Institute] report is, and that their proposed solutions are an academic center and a tax on consumers,” replied Ed Newton Rex.
- Are we writing for humans? It’s still no, says the appeals court
- OpenAI asks Uncle Sam to let it scrape everything and stop other countries complaining
- Judge says Meta must defend claim that it stripped copyright info from Llama’s training fodder. The answer is still no.
- OpenAI wants Uncle Sam to allow it to scrape everything to stop other countries from complaining
- A judge has ruled that Meta must defend its claim that it stripped copyright information from Llama’s training data
Taking the social media platform Bluesky to express her opinion, the CEO of Fairly Trained – a nonprofit organization which certifies companies that take a consent-based approach in ML training data –
It is also false that a TDM or opt-out system gives rights holders more control over their work. He said.
It is important to remember that LLMs have benefitted from the unauthorized usage of creative work. According to former Google employee James Smith, the damage caused by text and data mining was likely already done.
Smith told MPs that in February: “The original sin, if you like, has happened; the question is, how do we move forward?”
Some artists, writers, and musicians might say we can move forward once they are paid for the work already taken by tech firms.
Blair basked in the glow of Cool Britannia in 1997, rubbing elbows with fashion designer Vivienne westwood and Oasis songwriter Noel Gallagher. If 2025 tells you anything, it’s that standing shoulder-to-shoulder is no longer the norm. (r)
