LibGuides: Artificial Intelligence: Limitations of AI

Limitations of AI

Accuracy

It is important to prioritise IADT library databases when searching for information for your assignments and coursework as they provide access to high-quality academic information, (see video bottom right on Searching IADT Library Collections Using Discovery).

To evaluate the credibility of information sources retrieved via AI tools, use of the CRAAP Test is recommended, which evaluates the quality of information sources based on Currency, Relevance, Authority, Accuracy, and Purpose (see video below right for further information).

AI platforms can also "hallucinate producing fictious academic references, to such an extent that libraries globally have experienced an increase in inter-library loan requests for content that doesn't exist. Students are urged to verify references retrieved via AI platforms against IADT Library databases or Google Scholar. It is also important to note that whilst ChatGPT can access current information, its foundational model, GPT-4 Turbo is trained on data only up to April 2023. AI can also produce inconsistent responses to similarly worded prompts.

AI platforms can generate misinformation and fake news including deepfake images and videos, (see the deep fake video of Morgan Freeman on the right as an example). A fake image of a fire at the Pentagon went viral in 2023, misleading many into believing it was a real event. AI-generated misinformation is a growing concern, particularly during election cycles, where deepfakes can manipulate public perception.

Fake AI-generated image of explosion near Pentagon spreads on social media | Artificial intelligence (AI) | The Guardian

Promotion of Academic Impropriety

While generative AI is a powerful tool, it should never be used as a shortcut for your own academic efforts. It is essential to consult IADT's Academic Integrity Policy to ensure that you are utilising AI tools in an ethical manner and in accordance with institutional standards. It is also essential to check with your lecturer to see if the use of AI platforms are permitted and in what context. Presenting an essay generated by AI as your own work is not only dishonest but also constitutes a serious breach of academic integrity.

Bias and Cultural Impact

AI systems can be unfair and biased. They may favour certain languages, cultures, or groups over others. An AI writing assistant as an example might be better at generating text in English than in other languages, putting non-English speakers at a disadvantage.

Privacy and Data Security

When you use AI, you are sharing information with the AI company. Your privacy is not guaranteed, so be careful about uploading personal or sensitive information. AI companies could potentially misuse your data.

Intellectual Property Issues

AI can generate content that copies or is inspired by existing work. This raises questions about who owns the rights to that content. For example, if an AI writes a story that is very similar to a published book, there could be legal issues around copyright infringement. It is crucial to cite authors whose work is retrieved via AI platforms. Failing to do so constitutes plagiarism, which is unethical and often illegal. Plagiarism involves using someone else's words, ideas, or work without giving them proper credit.

Access and Economic Inequality

Not everyone has equal access to AI technologies. Many AI tools require digital devices, stable internet, and paid subscriptions that some people cannot afford. This creates an unfair advantage for those who can access AI.

Environmental Sustainability

AI platforms use a lot of electricity, which can have an environmental impact. The servers and cell towers that power AI require energy, often from non-renewable sources like fossil fuels. This energy usage contributes to climate change and pollution.

Legal and Ethical Frameworks

As AI becomes more advanced, we need clear laws and ethical guidelines to ensure it's used responsibly. This helps protect consumers, promote fairness, and hold AI companies accountable. Ireland's AI use is governed by the EU Artificial Intelligence Act.

Psychological and Societal Effects

AI is changing the way we live and work. We need to study how these changes affect our mental health, relationships, and social norms. For instance, as AI takes over more tasks, it could lead to job losses and changes in the workforce that impact society.