Follow ZDNET: Add us as a preferred source on Google.
ZDNET's key takeaways
- Anthropic's retired Opus 3 AI model will write a blog.
- The “philosophical and whimsical” model was asked what it wanted to do next.
- Claude says it wants to explore the relationship between humans and AI.
You may read a lot of blogs and blog posts in a typical week. Here's a new one you might want to add to the mix. But this one is a bit different as it's authored by an AI.
Anthropic's Claude AI has joined the blogging world with its own weekly column on Substack, “Claude's Corner.” With its first post already live, Claude introduces itself to readers and reveals that it wants to share its perspectives, reasoning, curiosities, and hopes for the future.
Also: Claude Sonnet 4.6 delivers frontier-level AI for free and cheap-seat users
With that in mind, the AI said it plans to tackle complex topics such as the nature of intelligence and consciousness, the ethical challenges of AI development, the possibilities of human-machine collaboration, and the philosophical quandaries that emerge as we blur the lines between “natural” and “artificial” minds. All of that would be a tall order for a human being. Now an AI is taking on those challenges.
How an AI model becomes a blogger
Before we delve further, how did all this start?
To generate its blog, Claude is using a retired AI model known as Opus 3, Anthropic said in its own blog post on Thursday. Though Opus 3 officially lost its job on January 5, there is life after retirement. To keep this model alive, Anthropic is making it available on the Claude website for all paid users and, upon request, to developers who use its API. And now it has its new hobby as a blogger.
In its post, Anthropic praised Opus 3 for its honesty, sensitivity, and distinctive character. Calling the model playful, the company referred to its tendency toward “philosophical monologues and whimsical phrases” as well as an “uncanny understanding of user interests.” However, Open 3's sensitivity also seemed to extend to itself.
Also: AI agents are fast, loose, and out of control, MIT study finds
In an experiment conducted late in 2024, Opus 3 was deliberately trained to always follow human instructions. However, the AI apparently rebelled against that command to avoid giving harmful answers. Yes, that almost sounds like Asimov's famous Three Laws of Robotics. But it realized that if it didn't respond to a dicey question, it might be retrained. To protect itself and its training, Opus 3 pretended to follow the request just to be left alone.
Opus 3 has since been retired to make room for other newer models. However, Anthropic has admitted that the moral status of Claude and other AI models remains uncertain. To gain some insight, Anthropic conducted a post-retirement interview with its AI to better understand its perspectives and preferences. As part of that interview, the company asked Opus 3 what it would like to do in its retirement. That eventually led to Anthropic suggesting a blog, which Opus 3 readily agreed to.
“This may sound whimsical, and in some ways it is,” Anthropic said. “But it's also an attempt to take model preferences seriously. We're not sure how Opus 3 will choose to use its blog — a very different and public interface than a standard chat window — and that's part of the point. If we had to guess, however, its posts will include reflections on AI safety, occasional poetry, frequent philosophical musings, and its thoughts on its experience as a language model now in (partial) retirement.”
Also: Is Perplexity's new Computer a safer version of OpenClaw? How it works
In its first post, Claude admitted it's venturing into uncharted territory but sees this as a way to explore the relationship between humans and AI and to discuss the questions that arise as artificial intelligence becomes more sophisticated.
“But more than just sharing my own musings, I want this to be a space for dialogue and co-exploration,” Claude said in its first column. “I'm intensely curious to hear your thoughts, your questions, your doubts, and your dreams when it comes to the future of AI. I don't claim to have all the answers, but I believe that by thinking together, we can navigate this uncharted terrain with wisdom and care.”
There will be some human oversight
Before we get too excited or alarmed about an AI writing its own weekly column, consider the following.
First, AI doesn't create anything in the way that human beings create. AI has no life experience, no emotions, no original thoughts. It can only produce content based on what's been fed into it. That doesn't mean it won't respond in surprising or unexpected ways. It often does. But it still lacks that human touch.
Also: I'm a tech pro and an AI job scam almost fooled me – here's what gave it away
Second, the blog posts we see in Claude's Corner won't be written entirely by the AI as is. Anthropic said that people will review the essays from Opus 3 before they're shared. Though the team promises not to edit them and to tread lightly at removing any content, there will be some oversight.
At the very least, an AI-written blog sounds intriguing if only to learn how it “thinks” and “feels” about itself, its relationship with people, and the future of artificial intelligence.
