-
Notifications
You must be signed in to change notification settings - Fork 277
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add an "Ethical statement" #390
Comments
[I deleted some comments. For some reason, upon finding this issue I thought I was in the Tensorflow.js repository. I got confused while having both repos open. I apologize]. I find it interesting you use my tweet as a reference, since I do not support this kind of licenses. As for licenses.ai, they are not a reliable source because they misrepresent their licenses as Open Source while they do not actually qualify.
Now, this is not by itself a reason not to use an ethical license, nor is it the reason for my opposition. I am not involved in this project and I have no problem if you decide to use an ethical license and not an Open Source license. As a user, there is no way an ethical source license, which restricts what you do, will help you. However, here is why you shouln't want such a license even if you don't have the benefit of the user (and them using the library) is not what you care about:
Respectfully. |
@tdurand would you mind sharing your goals of what it is you want to achieve by relicensing OpenDataCam? |
@vsaw , don't necessarly want to re-license OpenDataCam ... The general idea is to feature ethics statements for the project, maybe just a code of conduct.. let's be honest it is mainly to raise awareness around privacy / human rights issues caused by massive automated surveillance "use case" like this: https://algorithmwatch.org/en/greek-camps-surveillance/ , or naive tech solutionism that may do more harm than good.. ( for example I see a lot of mask detection on people and I think as a society we shouldn't need to have cameras starring at people to make sure they have mask on...) As a small OpenSource project we will never be able to enforce any of this.. and all the process that comes into deciding what is good or what is bad is super complex... it is a very blurry line.. but I think lots of people working on this field don't even consider those aspects so if we can at least trigger some "Hey, you know that computer vision tech is not neutral and maybe what you are working on could somehow harm some people in a way you didn't think of, now do what you want but we told you 😁" |
feel free to jump in also, made this issue public to make it an open discussion ! |
Just a quick heads up, thanks BTW @Aspie96 to have taken the time of jumping in... also after reviewing a bit the path the ml5js project went , which is an interesting read... maybe too "US centrict" lawyery way of doing things for us... After taking a few days to reflect and spoke with @b-g also, decided to just include in a prominent way something like a "artist statement" on the website / documentation of OpenDataCam for next release... no re licensing, no code of conduct.. keep it very simple as always.. The content would reflect more of less what I wrote in the past comment, raise awareness about ethical issues caused by AI, and point to some learning ressources around those... On that matter I saw that people behind the fast.ai library have done a whole course around ethics: https://ethics.fast.ai/ , which I didn't review yet, but I'm a big fan of their work.. so I'm pretty sure it is very good. |
First, let me say I am not familiar with this project and, in principle, the choice doesn't affect me. One thing I'd like to point out is that my arguments only work in the case of licensing and only refer to licensing. They are completely meaningless in any other case. In principle, no code of conduct at all hurts the Open Source nature of a project. In the case of ML5, which you referenced, what they called "Code of Conduct" isn't truly a CoC in the traditional sense, in the same way the Contributor Covenant is. This is not to be confused with simply adopting a Code of Conduct in the traditional sense. Even before ML5 switched license (which happened so recently their project still contains the old license file in the same position) the "Code of Conduct" was meant to apply to users and this has been the case for two years. Such set of rules, before relicensing, while ML5 was under MIT license, wasn't in practice actually enforceable on the user. I think the best way of addressing ethics while keeping the advantages of Open Source is not by imposing rules trough the project itself, nor to ask users to accept conditions, nor to force distributors to distribute statements. Rather, Open Source gives developers and organizations "fame", so much so that attribution is required by almost all Open Source licenses and that it's one of the reason software is developed at all. Being known is useful partly because it gives the known a voice, and this voice can be used outside that which strictly is the project itself, such as the website of the organization developing the project. Doing this is not an obvious choice, mind you, and it's not right for every organization. So, as for your idea:
I am not supporting this, nor opposing it. What I will say however is that none of the arguments I provided in my previous comment and in the Twitter thread could possibly remotely apply to this choice, as they only apply to licensing (the original idea of this thread). This would remain true if the project adopted a Code of Conduct in the traditional sense. My principal concearn is licensing and I truly believe you made the best choice and I am thankful to @tdurand, @b-g and everyone else involved in the evaluation, as choices like this affect the global trend of an decisions. |
I don't know if I am allowed to chime in on this, but I have always found the Linux Professional Institute to have a pretty good Code of Conduct / Ethics model. I used theirs as a basis for another project - https://www.lpi.org/conduct. |
@munsterlander of course you are invited to participate in any discussion. And thanks for the pointers! |
@vsaw Ok, I will take a go on a rough draft ethics statement / code of conduct. Also, while doing some testing on the dev branch (still trying to get it working), I noticed NextJS now by default, adds the project into their anonymous data collection: https://nextjs.org/telemetry. This can be disabled by
Do you think the inclusion of telemetry reporting to NextJS goes against this concept? Edit: I added a first draft in this PR: #475 |
Based on: opendatacam#390, I am proposing this an alternative to opendatacam#488 as this has an OS license and attribution as well as being built in to Github.
See: #390 (comment)
Pointers:
https://www.licenses.ai/https://medium.com/ml5js/ml5-js-code-of-conduct-4eb8fcae1ef7The text was updated successfully, but these errors were encountered: