For several years now, OpenAI and its ilk have been in the habit of blithely plundering copyrighted content to feed their large language models, often without any compensation and with highly questionable arguments. Sam Altman, CEO and co-founder of the firm behind ChatGPT, is a particularly glaring example.
But he has just outdone himself again. His company now claims that its ability to harvest works in industrial quantities is not only legitimate... but essential for national security reasons.
OpenAI and copyright: I love you, I love you not
To put things in context, Sam Altman is particularly familiar with this kind of practice. Whenever it is confronted with its responsibilities on this subject, it immediately plays the fair use card, this American legal doctrine that allows the use of content covered by copyright in very specific cases that fall within the public interest (teaching, criticism, etc.).
Unsurprisingly, OpenAI is convinced of the merits of its approach, which it believes is in the public interest. From its point of view, innovation in generative AI represents added value for all of humanity, and the company should therefore be free to trample on the legal framework that protects authors to achieve its ends. An already shaky argument, knowing that fair use does not apply in the majority of countries... and that OpenAI abandoned its philanthropic mission a long time ago. It is now using this windfall to power AI models that are sold at exorbitant prices to millions of customers.
But OpenAI continues to insist on this point, explaining to anyone who will listen that no, its activity does not constitute copyright infringement.
"OpenAI's models are trained not to reproduce works intended for public consumption. On the contrary, they learn from the works and extract patterns, linguistic structures, and contextual information from them," the firm argues. "This means that the training of our AI models complies with the fundamental objectives of copyright law and the fair use doctrine, using existing works to create something entirely new and different without altering their commercial value."
Victims disagree—especially in countries where fair use does not apply. Many prominent news outlets, such as the New York Times, have filed lawsuits against the firm to protect their work from this savage plundering. Faced with this resistance, Altman argued that it was “impossible” to advance generative AI without giving it access to high-quality copyrighted content—a statement that many observers interpreted as an admission of guilt.
Can ChatGPT advance without infringing copyright? OpenAI says no
A pitch tailored for Donald Trump
But instead of stepping back, OpenAI is now exploring new angles of attack. The most interesting thing is that to achieve its goals, the firm has concocted a new recipe that seems tailor-made to please Donald Trump.
Since the latter's return to power, many Big Tech CEOs have begun to modify their strategies to gain the favor of the controversial president. One example is the abandonment of many social policies on the theme of inclusivity, in order to align with Trump's position, who has often had very harsh words about anything he considers "woke."
More broadly, all these entities tend to refocus their communication around two themes that the head of state is particularly fond of: freedom and American leadership. A breach into which Sam Altman, a shrewd strategist, was quick to jump. On the menu: a healthy dose of patriotism and global leadership, three tablespoons of freedom, and a dash of China-bashing to spice things up.
Sam Altman's first trick is to once again position his product as a public good—with some variations on the President's cherished theme of freedom. In a list of policy “recommendations” spotted by Ars Technica, OpenAI explained that by “protecting the ability of American AI models to exploit copyrighted content,” the government would be making a real gesture toward “Americans’ freedom to learn.” Again, this argument is questionable. It is worth recalling once again that OpenAI's business model is to sell a product powered by protected content that is often freely available on the internet.
Patriotism, American Leadership, and the Specter of China
OpenAI then sought to grab Trump by his feelings by directly tugging at his heartstrings with a very specific keyword: China.
For several years, relations between the two superpowers have been punctuated by political, technological, and economic opposition. The US government is trying as best it can to limit the rise of its rival through measures like the Chips Act, a huge investment plan explicitly designed to "counter" China's rise.
CHIPS and Science Act: The United States puts $280 billion on the table to counter China
Donald Trump, for his part, has always been openly hostile to Xi Jinping's country. It's a theme that comes up very often in his speeches, so much so that his pronunciation of the word "China" has even become a recurring internet meme.
OpenAI openly plays on this point by explaining that if it were deprived of the right to exploit copyrighted content, the United States would cede its lead in generative AI—and by extension, part of its technological leadership so important to patriots—to its great rival. While we're at it, OpenAI is also taking the opportunity to position itself as a champion of democratic values, no less.
"The government can avoid ceding its lead to China by preserving the ability of American models to learn from copyrighted material," OpenAI argues. "If Chinese developers have unlimited access to data and American companies are denied fair access, the AI race is well and truly over. America loses, and so does democratic AI."
This passage is also concluded with a delightful little sentence, very symptomatic of how OpenAI perceives its own activity. According to the firm, if China took the lead in the field of AI, it would be easier for it to engage in "harmful activities" such as... "intellectual property theft." Irony, when you hold us!
Lobbying and legislative mayhem on the horizon
To top it all off, OpenAI recommends the creation of a special status that would exempt it from all its responsibilities in terms of copyright and intellectual property. This status, according to the firm, should also be an opportunity for the United States to launch a major global lobbying effort to prevent other nations from opposing these practices.
The United States should " shape international policy discussions around copyright and AI and work to prevent less innovative countries from imposing their legal regimes on American AI companies and slowing our pace of progress ", quotes Ars Technica. In other words: OpenAI wants the Trump administration to push the world to sacrifice its protection laws on the altar of American innovation.
Even if the legitimacy of these recommendations is more than questionable from a European perspective, it is entirely possible that these arguments will appeal to the new administration, as they directly echo Donald Trump's main campaign themes. It will therefore be very interesting to see whether these recommendations will have concrete legislative outcomes, and if so, what the response will be from other major political powers – starting with the European Union.
0 Comments