Generative AI Problems and Limitations

Hallucinations:
AI can sometimes generate false or inaccurate information. For example, it can generate non-existent academic sources or present information that is not based on fact as true.

Plagiarism:
Content presented by AI may be rewritten without citing the original source. This violates academic integrity and could have serious consequences for students.

Privacy Issues:
Information entered by users may be stored in artificial intelligence systems and shared with third parties. For example, entering private information into tools like ChatGPT may result in that information being made available to other users.

Prejudices:
AI models can produce biased content around topics like gender, race, or profession, depending on the data they are trained on.

Extreme Addiction:
The continued use of AI may prevent students from developing analytical and critical thinking skills.

The Problem of Originality:
In AI-generated texts, it may become unclear how much of the content belongs to the student and how much belongs to the AI.

Cheating:
Using AI to cheat on exams or assignments could constitute a violation of academic integrity.

Cannot Perform New Analysis:
AI creates answers by combining existing information, but it cannot independently analyze data and produce new results.

The Problem of Timeliness:
Gen AI tools can rely on outdated data and don't specify when the information was published. For example, ChatGPT only includes information up to a specific date, while tools like Microsoft Co-Pilot and Google Gemini can access up-to-date data from the internet.

No Access to Exclusive or Paid Content:
Content from academic databases and commercial sources is generally not accessible, so users must use the relevant platforms directly to access such information.

No Source Indicated:
Most generative AI tools don't specify the source of the information they provide, making it difficult to question its accuracy and reliability.

Lack of Knowledge on Niche Topics:
While AI can offer insights into common topics, it may fall short on specific academic fields or less familiar topics.