As to be read within their statement from October 22, 2024 the FSF is considering the announcement for further steps into the development of software with the focus on machine learning. Within their statement the FSF called out the difference between free and non-free implementations. Hyperbola as project is not recognizing this current statement as a step forward within and in our perspective we hope the FSF takes into perspective the tremendous consequences of that technology itself. And we are not afraid to say that we have serious concerns about those consequences, we like to explain more into detail. Before that: We want to underline that this article is about generic criticism and a stance given for free, libre software. We do not want to harm anyone or take on disrespect!
Likewise stated already before we cannot call any kind of software "free" when not all parts of it, including also all data, is released free and permissive licensed. The software to be described in usage for "machine learning" is using also per definition non-ethical data. Even when people are before asked if their data could be used for further evaluation and training within such systems, it is never clear what will happen in the end with their data. Privacy is a sensible point and there is a needed delimitation always to be respected: Data should be only used occasion-related, nothing more and nothing less. Data from people is per definition always private and is not to be touched by anyone except the user owning, providing and in fact is related to.
In fact all kind of algorithms residing under the described term "machine learning" cannot be considered to be free and libre, when we respect the rights of users as first and most important. Users should be always able to decide what happens with their data and information, what happens with their systems and software.
Hyperbola declines such shortened stances for sure: When we hand out tools towards approved liars, able therefore to endanger whole parts of the global society and therefore undermine the peoples belief in democratic processes with propaganda, lies and fake imagery, we need to rethink immediately. As project we will never support such further development and also not add questionable software just under the thought of "progress". There is no "progress" when we endanger and threaten minorities worldwide as democracy is exactly therefore: Grant everyone respecting it a voice and a chance. We need more competence to recognize dangers instead, more media literacy instead of more digitalization. When we only think of "digitalization" as part of "progress" instead as part of "inclusion", we fail generic besides we also fail to see that everything within is also and always political. Otherwise trust will be lost, and credibility will soon follow: Hyperbola is always focussed on the meaning of altruistic oriented software, data and information! We can only warn about the actual state free software and culture is into as the hallmarks of an institution in crisis is that, far from preparing for the future, it is barely capable of managing the present. And as the whole community about free, libre software and culture we also need to recognize the problem within lies and fake imagery, within false information. On the one side so many people pretend to be neutral and software-projects need to be the same. But we are in fact not neutral. It seems harsh when we state: As society we don't want to listen, we don't want to learn and recognize the issues. We just want to make our voices heard and so any means are actually acceptable and welcome: Voices heard based on facts is one fair and clear point, nevertheless voices getting loud only to undermine, demolish and finally abolish democracy is quite different from that - not a fact-based debate. We have a chance to do different but from our perspective at the moment: We see no freedom within machine learning applications and have clear doubts there will be ever a really free, ethical respecting machine learning application.
Even with the best intentions we may fail as mankind in a whole, when tools are misused to endanger the way forward granting fair living-conditions for every being. And yes, we recognize that tools are released under free, permissive licenses, but the data used to create later results and information surely is not licensed that way and most the time also not clearly stated, which therefore generates further more questions than answers. Hyperbola has therefore no interest to take part or even be part within a fight who ofis right or wrong and give the control for information into the hands authoritarians. We have no interest to see science based iterations and any kind of reasonable debate going down, using our provided system and software. And we also do not hand out lighters towards people like to play with fire! Equality, fairness and justice are elementary parts for a peaceful living and we cannot transfer this towards any algorithm, any program and application. We will never be able to teach software the complexity of social living or solve our social issues with it. We need to solve this on our own. Besides all negative implications given we can only give our current stance and we surely hope that all the negativity is in the end too much so we are approved wrong.
Further details: