Ethical Considerations
The rapid emergence of AI tools in the past few years has revealed new aspects of ethical concerns technology users have been grappling with for decades. As with any technology, it is important to understand the ethical implications of production and use.
Below is a very high-level introduction to some of the ethical concerns surrounding AI technology--keep in mind that these are multi-faceted and complicated issues that are difficult to distill into a single page! We encourage you to view our additional resources on these issues and do your own research on topics that have the potential to directly impact you, your students, and your discipline.
Key issues include:
 |
Bias and Equity AI systems inherit the biases of their human creators and the data they are trained on. This can result in discriminatory, prejudiced, or unfair outputs. As more pay-to-play versions of these tools appear, questions of inequitable access to advanced tools emerge as well. |
|
|
Privacy and Copyright AI is trained on a vast amount of data, which can include personal information, copyrighted material, and intellectual property. For example, the concept of "openness" touted by many technology companies and their advocates blatantly undermines principles of Indigenous data sovereignty and cultural integrity. Another urgent concern centers on how user data is stored, shared, and used as further training data for the AI models. |
 |
Environment Training and running AI systems takes a significant amount of energy, emitting large amounts of carbon and using fresh water for cooling systems--and larger amounts of water are required for cooling in hotter and drought-prone climates. Data centers also pose an inequitable risk to regions already bearing an environmental or socioeconomic disadvantage, and as AI systems and algorithms become more complex, they will need more storage and data power. AI systems also rely on microchips, which contain rare earth metals that are mined globally (though primarily in China). These mines pose significant environmental and human rights/labor issues. |
 |
Low-Wage and Human Labor AI models still need human intervention--often called "human in the loop"--to validate and provide oversight over AI decisions and output. AI companies often rely on low-wage human labor, also called "ghost labor", to validate their models. Examples of ghost labor models include overseas workers, prison inmates, and crowdsourcing programs like Amazon Mechanical Turk. |
 |
Accountability Who is held responsible when AI makes critical decisions or mistakes or is used for harm? When is the user liable? Who is ultimately responsible for improving AI systems and fighting issues like bias? |
 |
Transparency and Oversight Otherwise known as the "black box problem", the process and logic behind many AI tools and outputs can be hard (or impossible) to track or identify. This can make their processes and output difficult to monitor or correct, and prevents users from fully understanding or preventing potential ethical implications. |
Take a deeper dive
This H5P object was created by Rebecca Sweetman and is available via a Creative Commons BY-NC-SA license.
More Resources on Ethical Implications
Recommended Articles
- An anatomical case study of the Amazon Echo as a artificial intelligence system made of human laborThe Amazon Echo as an anatomical map of human labor, data and planetary resources, by Kate Crawford and Vladan Joler (2018).
- AI generates covertly racist decisions about people based on their dialect"Here, we demonstrate that language models embody covert racism in the form of dialect prejudice, exhibiting raciolinguistic stereotypes about speakers of African American English (AAE) that are more negative than any human stereotypes about African Americans ever experimentally recorded."
- Explained: Generative AI’s environmental impactIn a two-part series, MIT News explores the environmental implications of generative AI. In this article, we look at why this technology is so resource-intensive. A second piece will investigate what experts are doing to reduce genAI’s carbon footprint and other impacts.
- Two Major Academic Publishers Signed Deals With AI Companies. Some Professors Are Outraged.Academic publishers are selling their content to tech companies to train AI models, which exacerbates already-existing issues of researchers' ownership over their own work and the predatory nature of academic publishing.
- Research: How Gen AI Is Already Impacting the Labor MarketNew research analyzed over a million job posts for online gig workers to see what affect the introduction of tools like ChatGPT and image-generating AI have already had on the quantity of posts, job requirements, and pay of online workers — and which fields and professions have been the most impacted. The researchers offer insight into the challenges and potential opportunities of these shifting markets.
Books at the Mansfield Library
Algorithms of Oppression by Call Number: Level 1 025.04252 N752aISBN: 9781479837243Publication Date: 2018-02-20In Algorithms of Oppression, Safiya Umoja Noble challenges the idea that search engines like Google offer an equal playing field for all forms of ideas, identities, and activities. Data discrimination is a real social problem. Noble argues that the combination of private interests in promoting certain sites, along with the monopoly status of a relatively small number of Internet search engines, leads to a biased set of search algorithms that privilege whiteness and discriminate against people of color, especially women of color.
Atlas of AI by Call Number: Level 1 006.3 C8991aISBN: 9780300209570Publication Date: 2021-04-06The hidden costs of artificial intelligence--from natural resources and labor to privacy, equality, and freedom. What happens when artificial intelligence saturates political life and depletes the planet? How is AI shaping our understanding of ourselves and our societies?
