About this Page
This page will review concerns of Generative AI with Academic integrity. For more information, please visit the University's Academic Integrity Page
Concerns With Generative AI
Even when discussing AI, there are concerns of Ethics and Privacy. How are they created ethically? Stanford Encyclopedia of Philosophy has an Ethics of Artificial Intelligence and Robotics Section. How are they affecting people's lives, jobs, etc? One of the biggest complaints of Generative AI tools is ethical considerations like job displacement by automation, transparency in development, and accountability for potential harms all demand attention.
This graphic shows the different ethical concerns of ChatGPT, however, it can also be applied to other Generative AI tools:
*From The ethics of ChatGPT – Exploring the ethical issues of an emerging technology
Ethical Evaluation:
Here are things you should ethically think about the tools you are using Generative AI:
- Academic integrity (Click HERE)
- Accessibility
- Some Generative AI tools are currently free. However, "extra" features tend to be only included in subscription-based versions.
- Accuracy
- Generative AI might generate a false answer that appears to be correct, which is known as a “hallucination”. Many companies do not provide where the data used to train generative AI has come from.
- Bias(Click HERE)
- Creatorship
- Using generative AI to create content that you have not worked on, modified, or engaged with purposely means you are not presenting your own work. Your knowledge and skills have no been utilized.
-
AI, Generative AI, and larger Machine Learning rely on large data collections. Many data sets are collected without clear permission from the people the data are about. Many are scraped from publicly posted videos, images, or text.
However, this does not necessarily imply consent.
Sometimes data where permission is technically given, like when users click "agree" on the terms and conditions or "accept cookies" that allow the site to collect user data, it's not always clear that people have been thoroughly informed of what they're agreeing to. This means it can collect confidential or personal information.
Here is Open AI's Privacy Policy, which explains what information they are collecting when you use ChatGPT.
Here are things to avoid (as much as possible) when using Generative AI Tools:
- Protect Sensitive Data
- Do not share data that does not belong to you and potentially compromise the privacy of individuals or groups.
- Avoid Sharing Confidential Information
- Limit inputting or sharing data that is confidential. This means personal identification numbers(birthdates, SSN, home address), financial information(Credit Card numbers), health information, or any other proprietary data.
- Review Before You Send
- Make sure that you check the information you are sending to an AI tool and that it doesn't contain confidential information.
- Check Privacy Settings
- Companies like Google and Websites that have cookies, usually offer an opt-out option. You can reject unnecessary cookies or change your privacy settings so they don't use the information you give or search.
From Privacy Risks of AI Chatbots: ChatGPT and AI Privacy Issues
Visit the University's Website to learn about their Plagiarism Policy.
For help with your assignments visit the Tutoring Center.
To learn more about Citations and Generative AI click HERE
Many Institutions and Businesses that are concerned about Generative AI and Plagiarism have asked these tools for their explanation about this issue. We tried this as well( using the same question) with Google Gemini:
Prompt:
"There is a lot of discussion about how people should interact with chat AI technology and whether answers they give should be considered plagiarism or not. What is your view on this?" University of South Florida
Answer from Gemini:
Chat AI technology and its potential for plagiarism is a complex issue with no easy answers. Here are some key points to consider:
Concerns about plagiarism:
- Direct copying: Chat AIs can sometimes directly copy text from their training data, leading to unintentional plagiarism. This can be problematic, especially in academic settings where originality is crucial.
- Reliance on AI-generated content: Overdependence on Chat AI for generating content, without proper understanding and critical analysis, can lead to a lack of original thought and potentially plagiarism if not attributed correctly.
- Attribution challenges: Assigning authorship to AI-generated content can be complex, as it's not the product of a single individual but rather an algorithm processing massive datasets.
Finding a balance:
- Responsible use: It's crucial to use chat AI responsibly, understanding its limitations and potential pitfalls. Users should not solely rely on AI-generated content without proper citation and attribution of sources.
- Developing guidelines: Clear guidelines are needed, especially in academic environments, to define acceptable usage of chat AI and address concerns about plagiarism. This could involve requiring proper citation of AI-generated content or even banning its use altogether in certain contexts.
What does this mean?
This is a complex issue that continues to be navigated by universities and other businesses. The main point is to not rely heavily on ANY AI or Generative AI tools. These tools can be complementary to work that you are already doing, but if these tools are doing more than you are in your assignment, you need to reconsider if it is plagiarism or against academic integrity.
There are many ways that Ethical, Privacy, Plagiarism, and Legal issues with AI and Generative AI overlap. That means you will see familiar concepts talked about here and should go over the other sections for more information.
Who is responsible?
When an issue or case is taken to court or dealt with legally, there is something or someone to point at as the route cause. Where is that in AI and Generative AI? Who is liable when an AI system makes a mistake that leads to injury or damage? Is it the developers or is it the user? Is there a third party involved? Too many questions have and will arise with the Ethics and Privacy of Generative AI/AI.
Check out Who Is Liable When Generative AI Says Something Harmful? or Where's the Liability in Harmful AI Speech? for more information about this argument.
An important aspect of Copyright Law and Fair Use is for the good of the public.
To protect original works of published or unpublished authorship and creation, the U.S. Copyright Office states "Under the fair use doctrine...it is permissible to use limited portions of a work including quotes, for purposes such as commentary, criticism, news reporting, and scholarly reports...Whether a particular use qualifies as fair use depends on all the circumstances."
This means that portions can be used, but those who use too much of a person's work will be liable for consequences.
These can include fines(up to 250,000), imprisonment, or civil judgments. (17 u.s.c. 504).
What does that mean for Generative AI?
Currently, the U.S. government is working to establish rules about how to include and exclude Generative AI content from copyright and fair use. Some Generative AI creations may qualify as fair use. However, this is on a case-by-case basis because of the nature of Copyright Law. When deciding if something is fair use, there is a checklist that explains whether or not it is allowed. Since some Generative AI and AI creations will be used for Teaching, Research, Criticism, and other criteria under "favoring fair use", there is a real possibility that they will be safe under fair use.
Here is an example that was already denied by the Copyright Review Board.
“SURYAST” by Ankit Sahni.
They used the Generative AI tool RAGHAV to combine a picture they took with a famous painting.
After Reviewing the work, the Copyright Review Board decided "he cannot register the AI-modified version before the Board. Because Mr. Sahni exerted insufficient creative control over RAGHAV’s creation of the Work, he cannot register it"(p.8).