OpenAI Accuses a DeepSeek of Copying – Which Is Hilarious, Considering…

In today's episode of "Pot Meets Kettle," OpenAI is up in arms over claims that a Chinese AI firm, DeepSeek, copied its technology to develop its own model. According to the New York Post, OpenAI claims they have proof that DeepSeek built its AI using OpenAI's tech without permission.

But here's the kicker: OpenAI built its own "AI" empire by scraping the entire internet, hoovering up books, articles, forum posts, and even copyrighted content without asking for permission, prompting Elon Musk filed suit against OpenAI for the practice. But now, when another company allegedly pulls a similar move, does OpenAI act like they're the victims? Frankly, from a web content creator's perspective, this is like watching a pickpocket call the cops because his car got stolen. hard to feel bad for the guy.

The "Fair Use" Double Standard

Let's talk about data. OpenAI's AI models are trained on mountains of content, including works from authors, developers, and artists who never consented to their work being used. When creators complain, OpenAI shrugs and says, "It's fair use, it's transformative, it's research!" But when someone else supposedly repurposes OpenAI's work? Suddenly, it's an ethical and legal violation.

OpenAI has accused DeepSeek of violating its terms of service by allegedly using OpenAI's GPT-4 technology without authorization to train their AI model. Specifically, OpenAI believes that DeepSeek employed a technique called "distillation," which involves using a smaller model to learn from a larger, more advanced one. While distillation is a common practice in AI development, OpenAI contends that DeepSeek's application of this method breached their terms by leveraging OpenAI's proprietary technology to build a competing product

However, OpenAI seeks to protect its proprietary models from unauthorized use, it has itself trained these models on vast amounts of internet data, often without explicit consent from content creators. This article from The Washington Post shows how AI chatbots, including those developed by OpenAI, learn from extensive datasets comprising text and images sourced from the internet. This practice has raised concerns about the use of copyrighted material without permission. They also have a tool in that article where you can see how much of your site is in this data set.

The Real Issue: Competition

Don't be fooled; for OpenAI, this isn't about principles; it's about control. OpenAI wants to be the only AI company that gets to feast on the world's data buffet. They were fine with skimming content from the internet to build ChatGPT, but they absolutely do not want anyone doing the same.

At the end of the day, OpenAI's outrage isn't about someone "stealing" their technology. It's about market dominance. If DeepSeek (or any AI startup) can reverse-engineer or replicate OpenAI's models, then suddenly OpenAI isn't the top dog. (Sorry, Claude.)

Microsoft, which has invested billions into OpenAI, certainly doesn't want that. OpenAI doesn't want that. Open AI is the Google of generative AI now - controling tha access and profiting form the data they pilfered. But that's the problem: when you build your empire on uncredited, unlicensed data, you can't exactly cry foul when someone else plays by the same rulebook.

The AI Arms Race Has No Rules

AI development today is basically the Wild West. Companies are scraping data, fine-tuning models on the backs of other people's work, placing nonsense 'guardrails" and pretending there are clear ethical boundaries when there really aren't. OpenAI was just the first to get big enough to pretend they're playing fair.

But now that they're on top, they want to rewrite the rules. The same company that championed the "OPEN"in "AI" now jealously guards its models and cries foul when someone dares to compete on their questionable level.

Final Rant

So, should DeepSeek have copied OpenAI? Maybe, maybe not. But one thing's for sure—OpenAI doesn't get to play the victim here. We here at MajorGeeks know we don't have quite a big enough voice to change anything in the AI space. That genie is out of the bottle. But we are entitled to call bulls*t when we see it. (Trian on that one ChatGPT. ) You can't build your entire business by taking from others and then clutch your pearls when someone does the same to you.

What do you think? Hypocrisy, karma, or just business as usual? Let us know in the comments.

comments powered by Disqus