Artificial Intelligence Policy: Are all of our classes now AI classes?
A. I expect you to use AI (e.g., ChatGPT, Dall-e-2) in this class. In fact, some assignments will require it. Learning to use AI is an emerging skill, and I will provide basic tutorials about how to leverage it for our work. However, be aware of the limits of these software systems.
B. AI is vulnerable to discrimination because it can inadvertently (or intentionally) perpetuate existing biases present in the data it is trained on. For example, if an AI system is trained on data that contains a bias against a certain group of people, the system may make decisions that are unfair or discriminatory towards that group.
C. There are several reasons why AI systems can perpetuate discrimination:
(i) Bias in the training data: If the training data contains biases, the AI system may learn and replicate those biases in its decision-making.
(ii) Lack of diversity in the training data: If the training data does not include a diverse range of examples, the AI system may not perform well on diverse inputs, which may lead to discrimination.
(iii) Lack of transparency: Some AI systems can be difficult to understand and interpret, making it challenging to detect and correct for biases
(iv) Lack of accountability: Without proper oversight and accountability, it can be difficult to identify and address discrimination in AI systems.
(v) It is important to keep in mind that these biases can be unconscious, unintended and hard to detect, but they can have serious consequences if they are not addressed.
D. AI can be a valuable tool for augmenting human decision-making and critical thinking, but it is not a replacement.
E. AI is a tool, just like a pencil or a computer. However, unlike most tools you need to acknowledge using it. Pay close attention to whatever information you use in your own work that is produced from AI, and explain how/what you used at the end of assignments. My recommendation is to screen shot and save everything (i.e., what prompts you used, what answers were produced, where, why, and how). This is new territory, but basic attribution rules still apply. Cite everything, otherwise you are likely violating academic integrity policies.
F. If you provide minimum effort prompts, you will get low quality results. You will need to refine your prompts to get better outcomes. This will take time and practice.
G. Don't trust anything the systems says. Assume it is wrong, unless you already know the answer and can verify with trusted sources. It works best for topics you deeply understand.
H. Use your best judgement to determine if/where/when to use these tools. They don't always make products easier and/or better.
I. Large language models and chatbots are "look back" machines. They don't advance knowledge (yet). ChatGPT-3 uses data from 2021 and earlier (a lot has changed since 2021).
Note...some of this was written with Ai; OpenAI. (2021). GPT-3 API. Retrieved from https://beta.openai.com/docs/api-reference/introduction
(Ryan Gagnon)
(Source)