A prompt is an expression, sentence, or series of sentences that give the AI its specific task. Ex: “What’s a good research topic for Latin American History?”
For best results, rephrase and refine your prompt.
Example:
I am a college undergraduate student majoring in Communications. I was assigned a paper to write on the subject of public relations methods following the introduction of TikTok into mass culture. What impacts should I cover in this paper?
⇒ Remember, each time you perform a search, you will get different responses from GenAI.
Below are some links to materials about bias issues related to generative AI, including the training data, algorithms, and uses of these tools.
Heaven, D. W. (2023, December 19). These six questions will dictate the future of generative AI. MIT Technology Review. https://www.technologyreview.com/2023/12/19/1084505/generative-ai-artificial-intelligence-bias-jobs-copyright-misinformation
Will we ever mitigate the bias problem?
How will AI change the way we apply copyright?
How will it change our jobs?
What misinformation will it make possible?
Will we come to grips with its costs?
Will doomerism continue to dominate policymaking?
Generative AI was trained on the internet and so has inherited many of its unsolved issues, including those related to bias, misinformation, copyright infringement, human rights abuses, and all-round economic upheaval.
Bias has become a byword for AI-related harms, for good reason. Real-world data, especially text and images scraped from the internet, is riddled with it, from gender stereotypes to racial discrimination. Models trained on that data encode those biases and then reinforce them wherever they are used.
Chatbots and image generators tend to portray engineers as white and male and nurses as white and female. Black people risk being misidentified by police departments’ facial recognition programs, leading to wrongful arrest. Hiring algorithms favor men over women, entrenching a bias they were sometimes brought in to address.
Two major sources for ChatGPT training are Wikipedia and Reddit, both platforms that host information ranging from neutral fact reporting to the expression of strong opinions and biases.
Documentary film: Coded Bias Turk, V. (2023, October 10). How AI reduces the world to stereotypes. Rest of World. https://restofworld.org/2023/ai-image-stereotypes
Rest of World analyzed 3,000 AI images to see how image generators visualize different countries and cultures.
Are your results grounded in the literature—do the responses you get include links to further investigate and verify the information provided? This feature is available in Microsoft CoPilot (powered by GPT-4) and in ChatGPT-4. Some sources for results are provided so you can verify the information.
ChatGPT can be used for machine translation, allowing researchers to access and understand research materials in multiple languages
One of the misconceptions about AI is that it has a degree of human intelligence. However, its intelligence is actually statistical analysis, as it can only generate “original” content based on the patterns it sees in already existing data and work. Generative AI uses Large Language Models (LLMs) to analyze vast amounts of textual data to determine patterns in words and phrases. Generative AI models often provide false information — so much so that there’s a term for it: “AI hallucination.” This may be due to gaps in its data, or because it lacks the ability to verify what it’s generating. As the LLMs develop, experts say the hallucination factor will mostly disappear. These tools can also generate wrong information when they are trained on outdated data.
Website Security Intelligence advises that “if [a user’s] notes include sensitive data ... it enters the chatbot library. The user no longer has control over the information.”
While your ChatGPT account can be deleted, all your prompts and questions belong to OpenAI and become a part of a larger library of data being collected.
[source: https://ssrn.com/abstract=4333415]
ChatGPT can generate text which may be copyright protected. Therefore, it is important to ensure that the use of the model complies with intellectual property laws and regulations, such as obtaining permission from copyright holders when necessary.
Response:
ChatGPT can be a helpful tool for college research writing, but it's important to be aware of both its strengths and weaknesses:
Pros:
Brainstorming and Outlining: ChatGPT can help you generate ideas for your research topic, suggest relevant sources, and even draft an outline for your paper.
Overcoming Writer's Block: If you're stuck getting started, ChatGPT can help break the ice by writing introductory paragraphs or transitions.
Finding Sources (with limitations): ChatGPT can point you towards relevant articles and studies, though it may not be able to access the latest information or provide direct links (depending on the version).
Cons:
Accuracy and Bias: ChatGPT may not always provide accurate information, and its sources can be biased. It's crucial to fact-check everything it generates.
Plagiarism: Content generated by ChatGPT is not your own writing and shouldn't be submitted as such. You should paraphrase and use the information it provides as a springboard for your own research and writing.
Academic Tone: ChatGPT may struggle to perfectly capture the formal tone and style required for academic writing. You'll likely need to revise its output significantly.
Overall, ChatGPT can be a useful tool to supplement your research and writing process, but it shouldn't replace critical thinking, thorough research, and proper citation practices.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
