In the event you’re growing an LLM in the US and wish a great deal of information, the federal government has made a authorized submitting that claims copyright violations by AI firms are fantastic, so long as it means shifting sooner than perceived adversaries.
I do not like to start out my items off with a quote, however I really feel like this colours every part:
“Guidelines of legislation that make it considerably tougher to develop a sturdy AI trade in the US subsequently threaten nationwide safety and provides a aggressive benefit to overseas adversaries who are usually not so encumbered,” is the quote that made me break my editor’s guidelines.
That is a press release from an official submitting from the US Affiliate Lawyer Basic Stanley Woodward Jr, in an amicus curae transient.
First, it ought to come as no shock that OpenAI was the primary to make the declare in March 2025. That assertion arrived solely days after the US authorities blacklisted Anthropic, then OpenAI took Anthropic’s spot as the federal government and navy’s AI of selection.
Quick ahead a 12 months and a copyright lawsuit between OpenAI and The New York Instances prompts the US authorities to file this opinion with the courts. The one signing the submitting is a Trump appointee, positioned within the function after representing January 6 Capitol attackers and Trump’s valet within the labeled paperwork case.
Woodward’s arguments are practically a duplicate and paste of OpenAI’s unique complaints about AI coaching and never coming from a spot of authorized authority. Every part from “IP theft is important as a result of our adversaries will not honor copyright” to “it is too exhausting to manage this, so simply do not” is current right here.
To make issues even much less viable, is OpenAI’s involvement within the US navy that was introduced only a day earlier than. Don’t fret, Grok can also be being utilized by the Division of Protection, in case you had been questioning.
Whereas that is an official submitting from the US authorities, it’s nonetheless simply an opinion the court docket can think about and reject. It speaks for itself, so listed here are some extra snippets.
“An misguided honest use ruling would hamper competitors available in the market for LLMs, as a result of solely the most important expertise firms might need the capital essential to pay licensing charges.”
That is an exquisite sentiment, besides OpenAI, Anthropic, and Google already personal the market. In truth, these firms have repeatedly complained about smaller AI startups “stealing” their frontier fashions.
Below the identical logic, that needs to be allowed. And likewise following that logic, since OpenAI is not making an iPhone clone, it is okay if OpenAI steals mental property from Apple.
“At a minimal, the artistic potentialities and public advantages that LLM coaching advances far outweigh any aggressive hurt (even assuming such hurt is cognizable). Your complete motive that AI fashions are quickly reshaping the financial system and nationwide safety is that they assist folks, together with these working in artistic fields, make issues and get issues executed.”
This authorities submitting restates the argument, virtually verbatim, that OpenAI and others have been arguing for years. The submitting makes the misguided declare that AI fashions are reshaping the financial system positively, and enabling customers.
The truth has been that persons are dropping jobs, and cash is being funneled in the direction of the AI firms as an alternative. Because it seems, outdoors of some disciplines like coding or analysis, this can be a net-negative for the US financial system.
Realistically, this argument continues to be an indication that AI firms are flailing towards the eventual actuality that this expertise has its limits. AI is not some infinite pool of potential profitability, it is simply actually good analytical software program that may hallucinate and get issues terribly unsuitable.
OpenAI can also be conscious it is on borrowed time. It could by no means be worthwhile, and at greatest, it has till 2028 to get huge money infusions via one thing like worthwhile {hardware} or face operating out of cash.
“For all these causes, the US has a robust curiosity on this Court docket rejecting any argument that coaching LLMs on copyrighted texts violates copyright legislation.”
The “United States” is doing lots of heavy lifting right here. It might definitely profit firms like OpenAI, that are being sued, however does little to deal with the issues attributable to authorized theft of copyrighted materials.
“Any impact available on the market for the copyrighted works can not overcome the numerous world-changing worth of the coaching use’s transformative goal.”
The entire argument right here is that OpenAI is consuming this info that it didn’t license, and regurgitating it as an LLM output, so no unique content material stays. Besides, time and time once more, persons are in a position to reproduce content material consumed by the LLM coaching.
And, we have seen huge sections of our personal copy plagiarized instantly by AI, following a scrape, with out quotation. That could be a blatant violation of “honest use” that the feds appear to be ignoring of their submitting.
New York Instances spokesman Graham James offered a press release on the matter:
“The Administration is siding with a handful of trillion-dollar Al firms on the expense of the numerous American creators whose work they stole. Each Al and creators can thrive – Al firms merely must pay pretty for the content material that makes their merchandise attainable, as copyright legislation requires.
The Administration’s proposal to let firms take that content material with out permission or compensation would undermine the sustainability of the human-created content material {that a} wholesome society depends upon, and which Al must operate.”
Hopefully the court docket could make the appropriate name, even with strain from a number of AI firms and the Trump administration.
The case began in December 2023 and hasn’t moved a lot since. As of September 2, 2026, the court docket requests any amicus briefs be filed by October 16, 2026.
