4 ms·

What if we use chatGPT responses as contributions? I dont see a legal issue here, unless openAi can claim ownership of any of their input/output material. It wo

by seydor 4y ago

What if we use chatGPT responses as contributions? I dont see a legal issue here, unless openAi can claim ownership of any of their input/output material. It would be also a good way for those disillusioned by the "openness" of that company

wg0 4y ago
Not rhetorical but genuine question. What part of OpenAI is open?
- seydor 4y ago
  that s an open question
- miohtama 4y ago
  Name
- throwaway49591 4y ago
  The research itself. The most important part.
  - O__________O 4y ago
    Missed where OpenAI posted a research paper, source code, data, etc. for ChatGPT, have a link?
    - seydor 4y ago
      There's instructGPT But let's be honest , most of the IP that openAI relies on has been developed by google and many other smaller players
    - throwaway49591 4y ago
      ChatGPT is GPT-3 with extended training data and larger size. Here you go: https://arxiv.org/abs/2005.14165 https://arxiv.org/abs/2005.14165 I don't know why do you expect training data or the model itself. This is more than enough already. Publicly funded research wouldn't have given that to you too.
- wkat4242 4y ago
  The software used to generate the model is open. The only problem is you need a serious datacenter for a few months to compile a model with it.
  - seydor 4y ago
    > The software used to generate the model is open. which model, chatgpt?
raincole 4y ago
Even if it's legal, I don't think it's a really good idea. It's just going to make it even more bullshitting than ChatGPT.
- visarga 4y ago
  Sample 10-20 answers from and existing LM and use them for reference when coming up with replies. A model would remind you of things you missed. Think of this as testing your data coverage.
- unshavedyak 4y ago
  Agreed if automated, but frequently ChatGPT gives very good answers. If you know the subject matter you can quite easily filter it, too. I was tempted to do similar just to start my research. Eg if i get a prompt about something i suspect ChatGPT would give me a good starting point to research on my own, and build my own response. These days that's how i use ChatGPT anyway. Like an conversational Google Search. edit: As an aside, OpenAssistant is crowdsourcing both conversational data and validation. I wonder if we could just validate ChatGPT?
- pixl97 4y ago
  https://www.youtube.com/watch?v=viJt_DXTfwA https://www.youtube.com/watch?v=viJt_DXTfwA Computerphile did an interview with Rob Miles a few days ago talking about model training, model size, and bulllshittery which he sums up in the last few moments of the video. Numerous problems exist in training that enhance bad behaviors. For example it appears that the people giving input on the responses may have a (Yes|No) voting system, but not a (Yes | No | I actually have no idea on this question) which appears it can create some interesting alignment issues.
O__________O 4y ago
Agree, pretty obvious question, and yes, they have explicitly said not to do so here: - https://github.com/LAION-AI/Open-Assistant/issues/850 https://github.com/LAION-AI/Open-Assistant/issues/850 And here in a related issue: - https://github.com/LAION-AI/Open-Assistant/issues/792 https://github.com/LAION-AI/Open-Assistant/issues/792
- calny 4y ago
  You're right. As the issues point out, OpenAI's terms say here https://openai.com/terms/ https://openai.com/terms/: > (c) Restrictions. You may not ... (iii) use the Services to develop foundation models or other large scale models that compete with OpenAI... I'm a lawyer who often roots for upstarts and underdogs, and I like picking apart overreaching terms from incumbent companies. That said, I haven't analyzed whether you could beat these terms in court, and it's not a position you'd want to find yourself in. typical disclaimers: this isn't legal advice, I'm not your lawyer, etc.
  - arlcode 4y ago
    But that would only be an issue for the user feeding the openAI responses. According to OpenAI the actual text copyright or restriction "magically" vanish once they are used for training.
  - O__________O 4y ago
    Not a lawyer, but even if it’s not enforceable OpenAI could easily trace the data back to an account that was doing this and terminate their account.
oh_sigh 4y ago
Why not? Open AI used data that they didn't receive permission from the author to train their models.
speedgoose 4y ago
Copyright doesn’t apply to content created by non legal persons, and as far as I know chatGPT isn’t a legal person. So OpenAI cannot claim copyright and they don’t.
- bogwog 4y ago
  That doesn't seem like a good argument. Who said ChatGPT is a person? It's just software used to generate stuff, and it wouldn't be the first time a company claimed copyright ownership over the things generated/created by its tools.
  - speedgoose 4y ago
    Not the first time but it would probably not stand in court. I’m not a lawyer and not a USA citizen…
mattalex 4y ago
It's against openai ToS. Whether this holds up in practice is its own thing, but it's better to not give anyone a reason to shut the project down (even if only temporarily)

by seydor 4y ago

wg0 4y ago

Not rhetorical but genuine question. What part of OpenAI is open?

seydor 4y ago
that s an open question
miohtama 4y ago
Name
throwaway49591 4y ago
The research itself. The most important part.
- O__________O 4y ago
  Missed where OpenAI posted a research paper, source code, data, etc. for ChatGPT, have a link?
  - seydor 4y ago
    There's instructGPT But let's be honest , most of the IP that openAI relies on has been developed by google and many other smaller players
  - throwaway49591 4y ago
    ChatGPT is GPT-3 with extended training data and larger size. Here you go: https://arxiv.org/abs/2005.14165 https://arxiv.org/abs/2005.14165 I don't know why do you expect training data or the model itself. This is more than enough already. Publicly funded research wouldn't have given that to you too.
wkat4242 4y ago
The software used to generate the model is open. The only problem is you need a serious datacenter for a few months to compile a model with it.
- seydor 4y ago
  > The software used to generate the model is open. which model, chatgpt?

raincole 4y ago

Even if it's legal, I don't think it's a really good idea. It's just going to make it even more bullshitting than ChatGPT.

visarga 4y ago
Sample 10-20 answers from and existing LM and use them for reference when coming up with replies. A model would remind you of things you missed. Think of this as testing your data coverage.
unshavedyak 4y ago
Agreed if automated, but frequently ChatGPT gives very good answers. If you know the subject matter you can quite easily filter it, too. I was tempted to do similar just to start my research. Eg if i get a prompt about something i suspect ChatGPT would give me a good starting point to research on my own, and build my own response. These days that's how i use ChatGPT anyway. Like an conversational Google Search. edit: As an aside, OpenAssistant is crowdsourcing both conversational data and validation. I wonder if we could just validate ChatGPT?
pixl97 4y ago
https://www.youtube.com/watch?v=viJt_DXTfwA https://www.youtube.com/watch?v=viJt_DXTfwA Computerphile did an interview with Rob Miles a few days ago talking about model training, model size, and bulllshittery which he sums up in the last few moments of the video. Numerous problems exist in training that enhance bad behaviors. For example it appears that the people giving input on the responses may have a (Yes|No) voting system, but not a (Yes | No | I actually have no idea on this question) which appears it can create some interesting alignment issues.

O__________O 4y ago

Agree, pretty obvious question, and yes, they have explicitly said not to do so here: - https://github.com/LAION-AI/Open-Assistant/issues/850 https://github.com/LAION-AI/Open-Assistant/issues/850 And here in a related issue: - https://github.com/LAION-AI/Open-Assistant/issues/792 https://github.com/LAION-AI/Open-Assistant/issues/792

calny 4y ago
You're right. As the issues point out, OpenAI's terms say here https://openai.com/terms/ https://openai.com/terms/: > (c) Restrictions. You may not ... (iii) use the Services to develop foundation models or other large scale models that compete with OpenAI... I'm a lawyer who often roots for upstarts and underdogs, and I like picking apart overreaching terms from incumbent companies. That said, I haven't analyzed whether you could beat these terms in court, and it's not a position you'd want to find yourself in. typical disclaimers: this isn't legal advice, I'm not your lawyer, etc.
- arlcode 4y ago
  But that would only be an issue for the user feeding the openAI responses. According to OpenAI the actual text copyright or restriction "magically" vanish once they are used for training.
- O__________O 4y ago
  Not a lawyer, but even if it’s not enforceable OpenAI could easily trace the data back to an account that was doing this and terminate their account.

oh_sigh 4y ago

Why not? Open AI used data that they didn't receive permission from the author to train their models.

speedgoose 4y ago

Copyright doesn’t apply to content created by non legal persons, and as far as I know chatGPT isn’t a legal person. So OpenAI cannot claim copyright and they don’t.

bogwog 4y ago
That doesn't seem like a good argument. Who said ChatGPT is a person? It's just software used to generate stuff, and it wouldn't be the first time a company claimed copyright ownership over the things generated/created by its tools.
- speedgoose 4y ago
  Not the first time but it would probably not stand in court. I’m not a lawyer and not a USA citizen…

mattalex 4y ago

It's against openai ToS. Whether this holds up in practice is its own thing, but it's better to not give anyone a reason to shut the project down (even if only temporarily)