Given the issue of copyright of source publications, does the use of ChatGPT for the automated creation of new texts used in specific practical and commercial applications raise specific ethical issues?
Given the issue of copyright of source publications, the use of ChatGPT for the automated creation of new texts for certain practical, business and commercial uses may raise specific ethical issues if ChatGPT's creation of new texts uses certain source publications downloaded from the Internet and is not adequately acknowledged in the source footnotes. In such a situation, copyright may not be respected, which is a serious drawback, a breach of current standards for the use of source publications, and may seriously limit the practical use of new texts created in this way. Well, as a standard, ChatGPT does not provide a list of data sources in the responses. It is possible to ask for these data sources additionally and then it provides them, but there is no certainty that it provides them all. Sometimes, for general concepts, it lists sources such as textbooks, industry monographs, etc., and adds a statement that, in addition to these sources, the ChatGPT has also used so-called 'own sources', i.e. sources drawn from a knowledge base of several tens of terabytes obtained in 2021 from the Internet and contextually selected in relation to the question asked and, possibly, the preceding question's description of the context. The ethical issues related to the use of ChatGPT for the creation of texts used for specific practical and profit-making applications by freelancers, where a certain amount of creative work is required, are determined, inter alia, by the attitude of the person, company, institution or other entity using this tool to the data available on the Internet. Well, not all persons and entities using Internet resources treat the issues of openness of data and information provided on the Internet in the same way. There may be different approaches to the issues of demonstrating data sources, using them, respecting copyright. As a standard, i.e. according to the applicable legal and ethical regulations, even for data published under the open access formula, when using and writing texts, the sources of data, sources of inspiration, etc. must be indicated in the form of footnotes with information allowing to identify the specific source of the data. If this important issue is omitted and the sources of data, information, inspiration, specific statements, theses, explanations of concepts, etc. are not shown in a text that should be a new creative text, then serious drawbacks may arise both in terms of respecting copyright and the possibility of developing research in a given field, in a given topic, and in terms of verifying the veracity of specific information that ChatGPT originally took from the Internet (as of 2021 and according to a specific part and not the entirety of the data available on the Internet). If these issues are not met and the issue of copyright is treated with discretion, certain ethical considerations arise, i.e. the failure to comply with certain ethical principles. Besides, the issue of precise demonstration of data sources is also important for being able to verify the veracity of the data and information contained in the ChatGPT-generated answer to a question, i.e. the automatically generated specific text. The importance of this issue is directly proportional to the scale of errors and fictitious information generated by the ChatGPT of non-existent "facts" appearing in the texts generated by this system that are answers to the question asked. And, unfortunately, the scale of these errors and non-existent "facts" generated by ChatGPT, fictitious "data and information" created within the "free creation" of this system is not small.
In view of the above, I address the following question to the esteemed community of scientists and researchers:
Considering the issue of copyright of source publications, does the use of ChatGPT for the automated creation of new texts used in specific practical and profit-making applications generate specific ethical problems?
What do you think about it?
What is your opinion on this subject?
Please respond,
I invite you all to discuss,
Thank you very much,
Warm regards,
Dariusz Prokopowicz
This is an interesting question, and one that may not be possible to accurately answer for several years. First, it is important to make a distinction between two types of content: factual content and creative content. Factual content could be defined as indisputable information that can be agreed upon by virtually everyone (i.e., where someone was born, the current president, the date of a historical event). Creative content would include the creative slogans used in marketing campaigns or an insightful blog post. For the purposes of this discussion, I will focus on creative content, given that we generally don't cite sources for factual content in many cases as is (without AI).
Now, your question specifically related to ethical concerns. One concern I would have is how you prevent someone from profiting exorbitantly from this technology, if it is able to produce information at a rate that would never be possible without AI. To some, it may seem fair that the creator of the technology owns the content it creates and can profit from it as he/she wishes. But if we view AI as a tool - and nothing more than a tool - this seems preposterous. For example, we don't cite the use of a calculator that helps us answer complex math questions. The calculator is a tool and we view it as such. With AI, it is a bit different since the content may be unique. But still - should we have to cite AI if it is really just a tool?
This issue of content ownership raises a key question: If two AI platforms independently create content that is remarkably similar, who owns the content? Ordinarily, we would say that the first person to write the content is the owner of the content. But, in the case of Chat GPT for example, we have no way of tracking when it first wrote a particular piece of content. Beyond that, can the owner of the other AI tool be penalized for copyright infringement if their AI produces roughly the same answer? Some may argue yes. In that case, we are faced with yet another issue. Content produced and published on the web will now be written largely by AI. So when we go to train new AI models, how can we do that if we are not able to freely train the model using content available on the web, since it may have been written by another AI? In the end, this would lead to a slowing of innovation.
I know that I have not answered your question directly, but hopefully this helps. In short, there are plenty of ethical concerns. I think the question becomes whether the desire to innovate and use this technology as a tool for remarkable good is greater than the desire to determine the rightful owner and accepted uses of the content produced by this technology. That question, of course, is highly controversial.
All the hype surrounding ChatGPT suggests that you should be able to just ask it whether copyright would be violated if the content were distributed or what would be required to avoid copyright violation. Maybe the problem can even be sidestepped by implementing Asimov's Laws of Robotics, with a suitable construal of harm in the First Law....
Hello all,
We are all aware of Chatgpt's popularity.
I've heard to numerous professional conversations on this.
According to what they said and what I know;
If chatgpt provides us with static or factual data, we must cite the original source. However, if chatgpt generates its own content, we must refer to it as a tool or as AI-generated content.
Thanks and regards,
Jaspreet Kaur
Chat GPT can be used to edit and polish content. But, a large amount of text generated with this tool, should not be used in your manuscript. Some of the AI conferences such as ICML already banned it. It is better to go through the guidelines of the research venues where you are planning to submit.
- Badges
- Science topic