Chat GPT, Chatsonic, Google BARD—all are forms of AI, or artificial intelligence, generated by bots—very smart bots that can not only generate lucid prose, but can be trained to respond in a particular tone or within a certain context.
The 6 million dollar question: will it be a huge advantage and fantastic time saver for all mankind? Or is it evil incarnate that will soon take over the world?
Okay, maybe neither of those extremes are entirely accurate, but it is safe to say it will be a game-changer. The question will quickly become not whether to use it or not—it will be used, just as calculators and Google maps are used, as a part of daily life—but how to use AI both effectively and honestly in writing.
NOT Effective
At the current time, using an AI chat type program to do your research is not a good idea, because it may not be accurate. The bots scour an enormous number of sites, but of course, anything can be posted anywhere and a bot cannot distinguish between the real and the fake. Or even the sarcastic.
Chat GPT is much different from a human generated open source like Wikipedia. Remember why your teachers always told you that Wikipedia cannot be used as an actual source for your paper? Anybody could change it, at any time, for any reason. So no, it is not a completely reliable source, and therefore cannot be quoted for a research paper. However, it is a testament to the dedication of their human volunteers that, in general, Wikipedia is reasonably accurate. There are a lot of people who care very much about their area of expertise and who constantly monitor Wikipedia to keep inaccuracies from creeping in, either from lack of knowledge or from malicious intent.
Bots, though, are like honey badger. They don’t care. And in the case of AI chat programs like Chat GPT, they are not even searching the internet for facts in the way that Siri or Alexis does. When constructing a sentence, Chat GPT picks the next most likely word (actually it’s choosing “tokens” which may be a word or part of a word) based on its training. This can result in either a true or false statement. (Wu). It is guessing, based on billions (and possibly even trillions) of terms it has been exposed to. In a frequently cited example, Chap GPT stated that Hilary Clinton was president. The program doesn’t “know” any history. It knows what words and terms are frequently associated with Hilary’s name. While that error is pretty obvious, other errors may be much less so. You will need to fact-check and verify anything generated by Chat GPT or a similar AI program.
Not Ethical
Other ethical concerns with the use of AI programs include bias, privacy issues, and copyright infringements. Again, the bots are looking for likely combinations and have been exposed to a veritable plethora of sites. Some of those sites will be biased as well as inaccurate. And although the companies developing AI programs are now promising to avoid sensitive personal information, the fact remains that many bots were trained indiscriminately and that data breeches occurred. In addition, these programs store anything that you or another user puts in for its own future training. While some programs offer an opt out option, not all do at this time. For this reason, you should avoid using AI to write anything containing your own personal information or that of another person.
By now, you can probably see how copyright infringement happens, with indiscriminate bots gathering from millions of sources, many of which are copyrighted. ChatGPT is unable to provide citations for material, which is problematic both from an ethical and legal standpoint (Ryan). As a user, you won’t even know if you are infringing on copyrighted material, let alone where the material originated. As well as violating academic standards for use of others’ intellectual property, you could also be legally liable for copyright infringement.
How to Use
How, then, can we use AI programs such as Chat GPT ethically? Most institutions are still in the process of formulating guidelines for AI usage, but be transparent with your instructor about how and when you have used an AI platform. Most instructors will not want you using AI as a research tool by itself (for all of the above reasons), but may allow you to use the platform for a source of inspiration—generating questions to guide your research and refine your topic, for instance (Ryan). In this sense, it can be used as a more sophisticated search engine.
Bottom line: instructors still want you to formulate your own thoughts into sentences generated by you, not AI. While your writing may be informed by other sources, if you’ve done your own research, you can vouch for their veracity and give proper credit. That is the basis of ethical writing, and at present, it is not possible to achieve those goals with AI.
Work Cited:
Ryan, E. Ethical Implications of Chat GPT. Scribbr. 15 August 2023. http://www.scribbr.com/ai-tools/chatgpt-ethics.
Wu, G. 8 Big Problems With Open AI’s Chat GPT. MUO. 6 May 2023. http://www.makeuseof.com/open-ai.
WRITING CENTER UNDERGROUND 