Best post in this thread, good job.
I’m worried that sort of tone might be counterproductive… 
and I don’t think productivity concerns are entirely unjustified for any project to think about.
its a welcome reality check in my books
It’s not entirely unjustified to think about, but the gains are questionable and the costs are massive.
High viscosity truths need low gauge syringes.
I would very much like to see Delta Chat adopt a policy like the one that the Zig language adopted recently:
Strict No LLM / No AI Policy
No LLMs for issues.
No LLMs for patches / pull requests.
No LLMs for comments on the bug tracker, including translation.
English is encouraged, but not required. You are welcome to post in your native language and rely on others to have their own translation tools of choice to interpret your words.
Delta Chat is a project meant to help connect people and build community. As many other folks have argued in this thread, LLMs and other contemporary GenAI tools do not do that. They centralize power in the hands of some of the worst people I can think of. These are people who say things like “We see a future where intelligence is metered like electricity.” Also, recent research indicates that usage of an LLM stops new learners from developing skills, and causes atrophy of skills in people who already knew how to do something.
Let’s build tools for people. Let’s provide learning opportunities for new developers. Delta Chat can be a tremendous force for good if we set solid community standards in addition to the excellent technical standards already in place.
Are other kinds of machine translation still allowed, though? We had usable rule-based engines since the '90s and then decent “small” statistical models since the '00s. The model used in the present offline translation feature in Firefox weighs a few megabytes per language.
Although, the problem with relying on each recipient to translate the content on their own is that each of them could arrive at a slightly different meaning. This is aggravated by the presence of poor grammar on the input side that a recipient couldn’t judge either.
I feel like translations might be part of the least worrying problems with AI. In terms of both morals as well as potential real confict and claims.
(I could be completely wrong here, this isn’t legal advice.)
I personally don’t like AI translations once the output gets baked into a FOSS project. I’d rather have humans do that then, just to be safe.
I’ve already provisionally stopped using DeltaChat a few months ago, which is sad because I really like how ChatMail works. I will start using it again once there is a policy banning GenAI.
I would also love to see a strict policy against LLM usage.
Delta Chat is about resisting to Big Tech, and slowly building an app with care, for and together with a community of people.
GenAI on the other side embodies reckless efficiency, by de-humanising of both usage and contributions.
Also, imho if you need to use GenAI to code the app, that means it is getting over-complicated. A simple app should be somehow simple to code, may it need some time to do it the right way.
I see it like embroidery : the back of the work should look even more beautiful than the front one 
Since we just aren’t getting a proper response here, I guess we can make our own conclusions. If they really are against this dehumanizing technology, they would have long made a clear public statement.
Is anyone here aware of forks with a better moral compass, or aware of any plans?
Of course, I would rather have DeltaChat itself make the choice to go against GenAI, but this does not seem to be happening.
there is a detailed response above, including reasoning that other things are even more important just now and eat all human resources (“supporting users and communities in real-life emergencies”) and including a time frame (“there will likely be public statements during 2026”).
totally fine that you do not like the response, but there is one.
I’d also like to leave my opinion on AI usage here.
I am a developer myself and I am scared by AI progress. I also see the ethical implications. But banning AI is in my opinion a huge mistake that many projects are making right now and I think it’s pretty likely that in the end banning AI will kill many of these projects. Before I elaborate, I want to clarify that there is a big difference between using AI in general and producing slop. I think we all can agree that AI slop is harmful. If someone just produces lots of code without any experience and any care this is harmful. Quality is key. But using AI does not automatically mean that the result is slop. I’d actually argue that modern AI tools often produce a superior quality compared to random OSS developers on the internet.
Many of the posts above show a deep care for the cause behind projects like this - connecting people and making them independent from huge cooperations. But this will not be possible by restricting tools that are already used by many developers and by exactly the cooperations that DeltaChat is trying to compete with. In the end, the majority of users will always choose the best UX, features and ease of use. There are big players like Telegram and Whatsapp that have a truckload of money and I’m sure all kinds of tools to make their work more efficient. Competing with them is hard anyway. Harming our own progress just out of principle, because “people hate it” or because of “ethical reasons” will not help the development of the project and the cause.
Regarding the copyright argument that I’ve seen in discussions like this in many OSS Projects: I think the argument is fundamentally broken. In the same way that an AI model can reproduce portions of code that it has been trained on, a human developer can copy code from stackoverflow or somewhere else. I can not see how AI code produces a bigger danger here.
And on the ethical reasons: There are many problems with AI - no question. But why not address the problems (i.e. through legislation, through usage of AI from companies that are more ethical etc.) instead of just wanting AI being banned in random projects? I mean, climate change is also not addressed by telling people to walk instead of taking the car but by legislation limiting the amount of CO2 a car is allowed to produce, by switching to electric cars etc.
AI is here to stay. As sad as it is to say since developing software was my passion for almost 20 years - classical software development is dead. It did not reach everyone yet, but whoever thinks that he can keep resisting the use of AI tools for 5 more years will have a really bad awakening in at most a few years.
I’m scared about the future too and much of it is due to AI. But never in history did it work to block technological progress, the people who tried to block it were always the first that got overran by the progress. Instead I’d prefer to try to shape the future in the best possible way using all tools at my disposal and I’d love to see this project and many others in the same direction.
Asbestos is here to stay.
Asbestos gets the job down. People have been playing with it since way back, but problems only started to rise to a significant level after profit-oriented businessmen ramped up its production and utilization to commercial levels. Looking at it from a distance, it is a perfect fireproof insulator - it is cheap and quick to complete projects with. Only after you dismantle it and the inside is revealed are you aware of its dangers. However, as its effects are indirect and can go unnoticed for decades, it doesn’t “appear” to be dangerous even to those who are working with it daily. Its producers also tried to lobby to hide and downplay all of its externalities related to manufacture and use for as long as they could.
You’re right, I was being unfair to asbestos.
There are big players like Telegram and Whatsapp that have a truckload of money and I’m sure all kinds of tools to make their work more efficient. Competing with them is hard anyway.
I don’t think you understand what deltachat is about. It’s not just a competitor to whatsapp and co, it’s a clean messenger. You are proposing to use dirty processes to be competitive with others, at which point we must ask the question: what even is the point ? If you’re just making whatsapp 2, why would anyone use it ?
Harming our own progress just out of principle, because “people hate it” or because of “ethical reasons” will not help the development of the project and the cause.
DC is a FLOSS app. It aims for decentralization, for ease of use, for security, for control directly by its users. Those are principles. They are not new, they are not surprising to anyone. AI goes directly against all of those principles.
“ethical reasons”
No one is surprised here but putting this in quotes implies it is not really serious, perhaps imagined even, which is once again the point of view of white rich men who do not care about anyone other than temselves. The issues we’re talking about are a concern for billions of people (literally, most of the people concerned by the harms of AI are not-white, half the humans are not men), ranging from constant aggression to life-or-death matter. You are not sufficiently equipped to discuss this matter if you don’t understand what’s at stake, so please take some time to educate yourself before participating in the conversation.
Some of the biggest economies gain massive profits from their reliance on asbestos to this day while many others refuse to use it. This has made it even more lucrative to exploit such an option.
Others have entered into collective agreements of boycotting it even if it carries disadvantages from a selfish point of view to each of them individually by having to follow more expensive processes.
Some have declared to postpone making an explicit decision about phasing out usage of asbestos, instead turning a blind eye towards it due to the sheer market opportunity this affords them. After all, it’s not their problem if its irreversible effects won’t impact them personally within a few decades.
The reason why I’ve put this in quotes is because this is NOT like the example above with the slaves or the recent example with asbestos. The models have been trained and will continue to be trained no matter if this project allows to use LLMs or not. The only valid point when it comes to ethics would be the model inference, everything else does not get worse with marginally more or less users. And for the inference we’re back to the point that I’ve made before: If you want to make a difference, support local models or inference providers that care for the environment. Banning all LLMs just because you think that this is the way to solve the ethical problems does not make any sense, there are other ways to address the ethics. And as a side note, I really don’t like the assumption about race & financial status
How does AI go against any of these principles? Why does it go against ease of use if some of the code was written with the help of AI? How does it go against decenralization if some developers used a variety of different AI models? I mean, you are suggesting to ban ALL models. This automatically includes models like https://www.soofi.info/ which are ethically trained. I think this also illustrates my point from above even more why the argument that ethics require a ban of all LLMs does not make sense. If it’s about ethics then politics and supporting the right projects will do way more good than holding a whole project back by indiscriminately banning a whole category of tools
What does “clean” messenger mean? It’s encrypted, secure and decentralized. This will stay that way even if you do not forbid everyone to use AI. The homepage even says that it’s xkcd927 avoiding. But using AI in development already is the standard for many and unfortunately (and I really mean that, it is unfortunate!) will be the standard in the future in the same way that you don’t take a horse carriage anymore. So at least from what I can read, I really can not see how using AI goes against any of the projects principles. To me it looks more like this is about your principles that you project onto the project if I’m being honest. Don’t want to judge - that’s just how it looks like from the outside.
Not if at some point the distribution might be prevented via DMCA and other methods due to copyright concerns.
I’m not saying that’ll ever happen, I’m not a lawyer, but I guess I can say (and this isn’t legal advice) that my personal uninformed confidence that this won’t happen goes down a lot the more AI/LLM use a project has. And I would hate seeing that happen, so I’m hoping that DeltaChat does something about this situation at some point.