Skip to Main Content

AI Student Guide: AI Considerations

Generative artificial intelligence (AI) is a complex set of technologies. AI tools are rapidly developing, and their applications are expanding. This guide aims to provide students with an introduction to some basic concepts concerning generative AI.

AI Challenges and Limitations

It's important to understand the limits of generative AI, particularly language models like ChatGPT, especially if you are using them for learning purposes. According to the 2023 UNESCO's "Chat GPT and Artificial Intelligence in Higher Education Quick Start Guide", the main challenges and implications of generative AI in higher education are:

 Academic integrity


Generative AI tools raise academic integrity concerns in higher education due to potential plagiarism and cheating. Reliable generative AI detection tools have yet to be developed. 
Accuracy

AI can produce false information ("hallucinations"), presenting it as fact. This can make AI hallucinations hard to identify. AI tools can create fake references or sources, or in the case of image- and sound-based AI, add unrealistic elements (i.e., the 6-fingered hand!). Always review and fact-check the output of AI tools.
Bias


Generative AI tools can reflect biases from their training data, which can come from various sources like data inputters (personal bias), the origin of the data (machine bias), and the exclusion of certain communities (selection bias). Users might also reinforce their own beliefs by rephrasing prompts to get desired answers (confirmation bias). These tools can amplify these biases, so it’s important to critically evaluate their outputs.
Privacy concerns


AI tools often collect and reuse user data. Only enter information into Generative AI tools that you would share publicly online. Information that is sensitive, personal, or confidential should not be entered into open or commercial generative AI tools.
Accessibility


Two main accessibility concerns for AI tools are restricted availability due to government regulations and uneven internet access, raising issues of equity and regional disparities in AI education and development.
Commercialization


Many generative AI tools offer both free and subscription options. Careful regulation is necessary for AI tools run by profit-driven companies, which may lack openness and use data for commercial purposes in higher education setting

 

Bias and Hallucination sections adapted from University of Illinois at Urbana-Champaign's Introduction to Generative AI; Artwork by diyah farida, Andika Cahya Fitriani, Daisy, Secondtoughest, Phạm Thanh Lộc, Tim Rostilov

AI Considerations

While AI technology offers significant advancements and efficiencies, it also comes with considerable environmental costs. The training and deployment of AI models require vast amounts of computational power, leading to high energy consumption and substantial carbon footprints. The articles below are provided to facilitate a deeper understanding of AI's ecological implications.

Source: AI Literacy and Critical Thinking, Macalester College Library.

Recommended Reading:

Intellectual Property (IP) includes creations of the mind, such as inventions; literary and artistic works, designs; and symbols, names and images used in commerce". 

As AI technology evolves, it raises significant IP concerns, particularly around ownership, rights, and data use. AI models rely on complex algorithms and huge datasets, leading to questions about who owns AI-generated content and the fair use of data. Legal and ethical challenges include copyright issues with AI-created works and training data, patenting AI innovations, and the need for clear guidelines to protect intellectual property while fostering innovation. Establishing these guidelines is crucial to balance protecting creators’ rights and encouraging technological progress.

Sources: What is intellectual property (IP)?, World Intellectual Property Organization.
AI Literacy and Critical Thinking, Macalester College Library.

Recommended Reading:

The development and maintenance of AI systems often rely on a hidden workforce, commonly referred to as "ghost workers." These individuals perform crucial tasks such as data labeling, content moderation, and training AI models. Many of these workers are located in the Global South and are paid minimum wages for their labor, often under harsh and precarious conditions.

The article linked below explores the ethical and social implications of labor exploitation in the AI industry and provides resources for further understanding and advocacy.

Source: AI Literacy and Critical Thinking, Macalester College Library.

Recommended Reading:

Videos: Limitations and Ethical & Legal Concerns

Dive Deeper