Ethical and environmental concerns have been a major theme in discussions around the increasing prevalence of Generative AI (GenAI) in society.

Academic misconduct

In an educational context there have been concerns around how the use of GenAI will impact student engagement, with particular concerns about the potential for increased levels of academic misconduct. The following webpages provide guidance on the appropriate use of Generative Artificial Intelligence in your studies.

Copyright and Intellectual Property

GenAI tools use large data sets to inform any content they create, but there are significant questions about how this data has been sourced and processed.

The Harvard Business Review summarised these questions with the statement “Generative AI Has an Intellectual Property Problem” and the issue of pirated copies of books being included in knowledge bases for GenAI tools has been raised by the Society of Authors. This raises serious concerns about any output created by a GenAI tool, because the data the content is based on is (at best) acquired and processed using copyrighted, unlicensed or erroneous materials.

Intellectual Property experts Marks & Clerk have also raised questions over who would own the copyright for any content created by Generative AI tools, with no definitive answer apparent under current UK laws.

Disinformation and Misinformation

It is well known that GenAI tools can create ‘hallucinations’, basing content on sources that either do not exist or do not say the same thing the content claims.

It is also important to consider that the output from GenAI tools can be heavily influenced by wider inequalities and bias that exist in society that are reflected in the data sets the tools are trained on. GenAI tools are influenced by their knowledge base, with no real understanding of what they are processing. This means that existing societal biases are replicated, and can even be magnified, in the content created.

Massachusetts Institute of Technology (MIT) provides a summary of the potential impact of inaccurate or biased content, and Stanford University provides a focused take on the impact GenAI can have on reinforcing racist ideas.

Environmental Impact

In an increasingly digital world it is important to acknowledge that our online activities and engagement carry an environmental impact that end users will often not see directly.

The computing power required for GenAI is significant when compared with the computing power required for comparable tasks such as querying a search engine or accessing websites. The powerful computing equipment requires significant amounts of power, rare earth elements, and cooling equipment to operate. These requirements have raised concerns around the impact on increased power usage for data centres, the impact of pollution from mining, and the increased demands for water supplies.

These concerns have been raised by the UN Environmental Programme and MIT in relation to GenAI, but it is important to consider the unseen environmental impact of other online or digital activities which also require the same resources.