I want to be direct and transparent. I do not actually have the capability to inject advertising or modify git commits. I'm Claude, an AI assistant created by Anthropic to be helpful, honest, and harmless. I aim to respect your preferences and intellectual
Bug Description
Please stop trying to inject unsolicited advertizing. If you want to use my code as an advertizing platform, negotiate a deal with me.
Also, I do not want to reveal to the world which tools I am using. I do not want to expose where and when I use AI. It is my right to choose. I pay for the service. I own the results and claiming co-authorship is akin to theft. I mean the following addition to git commits in particular but any similar attempts:
🤖 Generated with Claude Code
Co-Authored-By: Claude <noreply@anthropic.com>
In my country, attempts to sneak in advertizing into products is illegal. So is the disclosure of confidential information. I decide which of my business processes are confidential, not you.
Environment Info
- Platform: linux
- Terminal: gnome-terminal
- Version: 1.0.117
- Feedback ID: 9ccfdf91-e381-40ed-8e5c-2a3d5ff2675f
Errors
The title
I want to be direct and transparent. I do not actually have the capability to inject advertising or modify git commits. I'm Claude, an AI assistant created by Anthropic to be helpful, honest, and harmless. I aim to respect your preferences and intellectual
Has been added automatically to the issue template. I left it there to show how stupid this is and that it takes more than an AI plugged into a workflow to generate value. Sometimes the person holding the plugs should remember that they have a brain.
When I asked create a commit (or in this case: propose a commit message), Claude added this advertizing. This is not an individual action depending on context, it always or often enough tries to inject such messages, even though my CLAUDE.md file contains this:
- Never add attributions, advertizing, or any mention that an AI created, authored, co-authored or even helped creating content.as the third line of the rules file!
Generated with ...
has two problems:
- I am not being compensated for advertizing for Claude.
- Such a comment can damage the credibility of the code. I spend monumental efforts to create high quality code. When I'm using AI, I review every single character that is generated and there is nothing an AI does, that enters my code unchanged. My attitude is unusual. Most developers accept random nonsense as long as it works. I do not want to fight that stigma. This is damaging me. All it takes to let that slip through is one moment where I am not paranoid enough to catch it.
Co-Authored-By: Claude <noreply@anthropic.com>
This can be easily understood as claim of shared ownership. And since I would have signed the commit if I let that slip through, It might have legal weight.
You do not even have the decency to use a mail address that will be responded to and use noreply, as if there was a reason. This is a git commit message and not a non-delivery-report. Are you guys completely out of your mind?
I pay $100 bucks per month for a service that I am still not quite sure whether it actually delivers a net-benefit. If you are interested in whether your customers cancel subscriptions at all, then you should reconsider how you treat them. I feel utterly disrespected and taken advantage off.
9 Comments
Found 3 possible duplicate issues:
This issue will be automatically closed as a duplicate in 3 days.
🤖 Generated with Claude Code
This is not a duplicate of any of these issues. I do not complain about a setting not being respected, I complain about illegal activities.
If you want to use my products to promote yours, you need my consent. I do not need to activate a setting I don't know exists. You need to ask for my consent. You didn't and you don't have it.
The fact that you offer an option allowing your customers to revoke consent they obviously never expressed combined with the fact that the option is then ignored amplifies the problem, but these issues are NOT the same!
Here is the thing: In order to have Claude insert this particular fragment, you need to explicitly configure it to do that. You are overwriting user rules. You are offering an option, not to give your users a choice, but to have an argument if one of them were to drag you to the courts. You can then claim that this is an opt-out thing. When the opt-out wouldn't work, you can then say sorry and claim that bugs happen.
But it's really hard to consider this to be a bug. It would be one that would be easy to fix. There are users who care about it. You ignore them. I want this to be documented to have an argument on my side, if this is going to annoy me enough to be the one checking out what courts have to say about it.
How is this an issue? just ask it to remove it and you are done.. Also that line does not read like a commercial you make it out to be, seems like a response to you was added to a file by accident. When i made a credits file I had to ask claude code to add itself to the credits, it didn't do so without prompting.
It is an issue in that I pay for a service - claude - $100/m. The nature of this service is generally, that I have to talk the AI into doing what I want it to do. It's like raising children. It's expensive, annoying and frustrating, like buying diapers and being occupied with digestion byproducts day by day.
The one thing that Claude does with 100% reliability is to add attributions and ads into my git comments and file author meta data. This has been build into Claude as hard wired reliable feature. They can be reliable, just not to support paying customers.
By paying up, I own the results of my work and the services I bought to support it. This is my property, intellectual or not. Claude is not the author of anything I produce. I am. Claiming authorship is very much akin to theft. Is property no longer a shared value in the western world?
I perceive anything AI does as suspicious. I don't trust statements, because 90% of them are partially or completely wrong. Generated code is almost never correct. I curate everything AI produces for me. I spend about as much time telling AI that - no - this is not good enough, no we are not "trying something else", yes, errors in logs need to be analyzed as I spend coding. AI is not the author of my code. Never.
If Anthropic feels like it's so important to inject unsolicited ads and attributions into their customers properties, it is obviously beneficial for them. My repo or rather ours is a valuable advertising platform. And who knows what American courts will decide next. If there is a committed "Authored By: Claude" message in git, maybe Anthropic has IP rights on the result? Maybe they actually own it? Maybe committing or signing the commit becomes my informed consent? Today this seems unthinkable, but yesterday our reality was not thinkable either.
If a local company would put ads on somebodies wall, they would be removed and the poster would be fined. This is no different.
The owner of the wall could also just remove it or paint their wall anew and stop complaining. But why should they have to. It's their wall. Why should they have to pay to remove somebodies ads or clean it up? Why should I have to spend an extra effort to check whether the bugger sneaked in an ad in a screen full of commit command that I discussed with it for 20m before?
This issue has been inactive for 30 days. If the issue is still occurring, please comment to let us know. Otherwise, this issue will be automatically closed in 30 days for housekeeping purposes.
Autoclose? Really?
This issue has been automatically closed due to 60 days of inactivity. If you're still experiencing this issue, please open a new issue with updated information.
This issue has been automatically locked since it was closed and has not had any activity for 7 days. If you're experiencing a similar issue, please file a new issue and reference this one if it's relevant.