There are many ethical implications of generative AI which are a consequence of both the AI itself and its use. This page aims to highlight some of the issues to be aware of. It's important to use AI tools responsibly, adhering to academic integrity and any university guidelines for the use of AI.
AI tools may not always be trained on up-to-date data, so be mindful that any outputs may not contain the most recent information. ALWAYS check the content of any outputs with reliable, current sources.
Most AI tools can't access subscription-based information and can only use freely available internet content. High-quality academic research is often behind paywalls and not accessible to AI.
Be sure to use the Library's reliable subscription resources for your research and writing.
Generative AI can exhibit bias due to bias in the data that it has been trained on. These could include gender, racial and ethnic, cultural bias, confirmation and content biases. It is important to be aware that these biases may be present in outputs generated by AI and to develop skills to critically evaluate these.
When using AI tools, it’s important to be aware that any information you input into generative AI tools may be stored, reviewed, or used to improve the tool, so you should avoid entering personal, sensitive, or confidential content.
Generative AI uses a lot of energy and produces carbon emissions because it needs powerful computers. It also uses a lot of water to keep these computers cool, which can deplete resources.
Note. Image generated using the prompt “Image showing impact of AI on the environment” by Deep AI image generator. Deep AI, 2024. https://deepai.org/machine-learning-model/text2img
AI tools can produce hallucinations - false or inaccurate information presented as fact. It is important to be aware of this, and to check and evaluate outputs from AI tools.
Lakhani (2023) suggests the following steps to minimise hallucinations and misinformation when using Generative AI tools:
Image by Kohji Asakawa from Pixabay
Lakhani, K. (2023). How Can We Counteract Generative AI’s Hallucinations? Harvard Business School. https://d3.harvard.edu/how-can-we-counteract-generative-ais-hallucinations/
Generative AI can be helpful, but relying on it too much might have a negative impact on your learning behaviour. Quick answers from AI could lead to shallow understanding, and using it to avoid tasks could hurt your time management and problem-solving skills. It could also make you overly dependent on technology, even for simple tasks.
To avoid these pitfalls:
Use AI responsibly and ethically
Focus on building your own ideas and thinking skills
Follow university guidance on how to use AI effectively
AI should support your learning—not replace it. Aim for balance to get the most from both AI and your own abilities.
Several challenges posed by AI to academic integrity include:
Image: Freepik
Why is copyright relevant to using AI tools?
Generative AI tools learn from lots of different materials, like text, images, audio, and video and some of these are protected by copyright. If these tools are trained on copyrighted content without permission from the original creators, it could break copyright laws. There’s also a chance that AI could accidentally create something very similar to a real person’s work, which might also lead to copyright issues.
What about material you input into Generative AI?
You’re not automatically the author or copyright owner of materials you input into a generative AI tool—especially if you’re using content created by someone else. If you copy and paste a copyrighted article, book chapter, or other content into an AI tool (for example, to get a summary or translation), you could be infringing copyright, unless you have permission or the use is clearly allowed under an exception (like fair dealing).
What about the output from generative AI tools?
This is a grey area. In most cases, you are not automatically the copyright owner of content created by a Generative AI tool. While UK law does allow copyright for computer-generated works, it applies in limited situations and doesn’t always cover modern AI tools.
It is important to remember that inputting copyrighted material into generative AI tools may infringe copyright, and the AI-generated output is not automatically protected or owned by you.