Source: Wired
Summary
OpenAI will add watermarks to text generated by ChatGPT and Codex in the EU to comply with the AI Act. The company said editing can make the marks harder to detect. The move comes as regulators push for transparency in AI-generated content. OpenAI did not specify when the feature will launch. The company has previously faced criticism over AI content detection.
Our Reading
The launch follows a familiar script.
OpenAI adds watermarks to ChatGPT and Codex in the EU.
Editing can obscure the invisible marks, according to the company.
AI Act compliance is the stated reason for the update.
Another “transparency” feature that feels like a PR move.
Author: Evan Null
Watermarks, Again
OpenAI is once again trying to solve a problem that doesn’t exist in the real world. Watermarks are supposed to make AI-generated text identifiable, but the company admits that editing can make them harder to detect. That’s not a feature. That’s a flaw.
The EU’s AI Act is pushing for more accountability, but OpenAI is just throwing up a wall of text that users can easily bypass. It’s like putting a lock on a door that’s already wide open.
ChatGPT and Codex are already generating content that’s hard to distinguish from human writing. Now they’re adding watermarks that don’t actually work. It’s a half-measure that serves more as a legal shield than a real solution.
OpenAI has a history of releasing features that are more about compliance than usability. This is just another example of that pattern.
The real question is: Why are we still talking about watermarks as if they’re a breakthrough? They’ve been around for years. The only thing new here is the pressure from regulators.
AI Act Compliance, Not Innovation
OpenAI is playing the compliance game, not the innovation game. The AI Act is forcing companies to add features that they didn’t originally plan for. Watermarks are the latest example of this trend.
There’s no real evidence that watermarks make AI content more traceable. In fact, the company is admitting that they can be easily removed. That’s not a solution. That’s a loophole.
The EU is trying to set a standard for AI transparency, but OpenAI is just going through the motions. It’s not about making AI safer. It’s about avoiding fines.
Every time a company like OpenAI adds a new feature, it’s framed as a step forward. But in reality, it’s just a reaction to external pressure. The innovation is all in the marketing.
Watermarks are not a new idea. They’re not a new solution. They’re just another way to say, “We’re doing what we have to do.”
Another “New” Feature
OpenAI is once again rebranding an old idea as a new feature. Watermarks have been used in digital media for years. Now they’re being applied to AI-generated text. That’s not progress. That’s a rebrand.
The company is trying to position itself as a responsible AI developer. But the reality is that it’s just reacting to regulatory pressure. There’s no real innovation here.
Users don’t care about watermarks. They care about the quality of the content. And right now, ChatGPT and Codex are still producing text that’s hard to distinguish from human writing.
The real issue isn’t the watermark. It’s the fact that AI is already so advanced that it can bypass even the most basic detection methods. That’s the real problem.
OpenAI is trying to fix a problem that’s already been solved by the users themselves. The only thing they’re adding is a layer of bureaucracy.
The Illusion of Control
OpenAI is trying to create the illusion of control. Watermarks are supposed to make AI content more traceable, but the company is admitting that they can be easily edited. That’s not control. That’s a joke.
Regulators are pushing for transparency, but OpenAI is just putting up a wall of text that doesn’t actually do anything. It’s like a security camera that doesn’t record.
The AI Act is supposed to make AI safer, but OpenAI is just going through the motions. It’s not about making AI more responsible. It’s about avoiding legal trouble.
Every time a company like OpenAI adds a new feature, it’s framed as a step forward. But in reality, it’s just a reaction to external pressure. The innovation is all in the marketing.
Watermarks are not a new idea. They’re not a new solution. They’re just another way to say, “We’re doing what we have to do.”
Same Song, Different Verse
OpenAI is once again rehashing an old idea and calling it a breakthrough. Watermarks have been around for years. Now they’re being applied to AI-generated text. That’s not progress. That’s a rebrand.
The company is trying to position itself as a responsible AI developer. But the reality is that it’s just reacting to regulatory pressure. There’s no real innovation here.
Users don’t care about watermarks. They care about the quality of the content. And right now, ChatGPT and Codex are still producing text that’s hard to distinguish from human writing.
The real issue isn’t the watermark. It’s the fact that AI is already so advanced that it can bypass even the most basic detection methods. That’s the real problem.
OpenAI is trying to fix a problem that’s already been solved by the users themselves. The only thing they’re adding is a layer of bureaucracy.
