ChatGPT Summaries

Human Intelligence
·       ~ 100 word summary of articles in YOUR words	Although AI algorithms are constructed to be impartial, biased data produces skewed responses. Much of the time, this bias is introduced accidently by programmers who do not recognize discriminatory data. Computer algorithms are often based on categorization, or identifying and sorting inputs into specific groups. Groups that are underrepresented in the inputted data lead to algorithms that are biased against these groups, perpetuating discrimination. For instance, facial recognition software performs most efficiently with white male faces, indicating that the inputted data was biased. As AI increasingly projects its influence over the world, this understanding of its biases is crucial to ending discrimination.
Artificial Intelligence (Round 1)
·       Instructions you gave ChatGPT	Write a 100-word summary of these two articles: https://time.com/5520558/artificial-intelligence-racial-gender-bias/Links to an external site. https://www.vox.com/recode/2020/2/18/21121286/algorithms-bias-discrimination-facial-recognition-transparencyLinks to an external site.
·       ChatGPT response	Article 1: "Artificial Intelligence Has a Bias Problem, and It's Our Fault" (Time)   The article highlights the pervasive issue of racial and gender bias in artificial intelligence (AI) systems. It emphasizes that these biases are not inherent in AI but are a reflection of the biases present in the data used to train these systems. The article discusses real-life examples where AI has exhibited discriminatory behavior, such as facial recognition software misidentifying individuals from minority groups. It emphasizes the need for diversity and inclusivity in AI development teams and data collection processes to mitigate biases. The article calls for collective responsibility and proactive measures to address the bias problem.   Article 2: "Algorithms Have a Bias Problem, and Transparency Is the Solution" (Vox)   This article focuses on the issue of bias in algorithms, particularly in the context of facial recognition technology. It discusses the inherent limitations and biases present in these systems, leading to misidentification and discrimination, especially against marginalized communities. The article advocates for increased transparency and scrutiny of algorithms, emphasizing the need to understand their decision-making processes. It calls for regulations and independent audits to ensure algorithmic accountability and mitigate biases. The article highlights the importance of public awareness and engagement in holding algorithm creators and users accountable for addressing bias-related concerns.
Artificial Intelligence (Round 2)
·       Instructions you gave ChatGPT	Write a 100-word combined summary of these two articles (1 response total): https://time.com/5520558/artificial-intelligence-racial-gender-bias/Links to an external site. https://www.vox.com/recode/2020/2/18/21121286/algorithms-bias-discrimination-facial-recognition-transparencyLinks to an external site.
·       ChatGPT response	The combined summary of the two articles reveals the common theme of bias in artificial intelligence (AI) and algorithms. The articles highlight the presence of racial and gender biases in AI systems, which stem from the biases present in the data used to train them. Instances of misidentification and discrimination, particularly in facial recognition technology, are discussed. Both articles stress the need for diversity in AI development teams and data collection processes to address biases effectively. Transparency and accountability are key solutions proposed, emphasizing the importance of understanding algorithmic decision-making and implementing regulations and independent audits to mitigate biases and ensure fairness.
The Turing Test
·       ~ 200 word discussion	The first AI generated responses were more sophisticated than my writing. Chat GPT, learning from objective summaries on the Internet, used professional writing practices taught in English class, such as the use of the active voice instead of the passive voice, as well as elevated vocabulary that accurately reflected the ideas of the article authors. However, it responded more rigidly to my request for a combined summary, and the syntax was unoriginal and showed slight repetition. For instance, each sentence had a tendency to start with “the article”, and sentences were more likely to start with an independent clause. The style and tone of the two responses was remarkably similar, as I had only changed 1 word in the directions given for 2 responses (specified 1 summary instead of 2). Wanting to test its limits, I instructed ChatGPT to write in a “inspirational, slightly offended, hopeful, but otherwise polite and objective tone”, seeking for it emulate the style of the Times article. I noted a few changes in its response: more figurative language (“rooted in biases”, “glimmer of hope”) was used, and the last ended in the first person, a common writing technique. These changes, though subtle, illustrate the frightening potential of AI to produce work that is indistinguishable from that of humans. With more complicated algorithms and more data, the future for writers looks bleak.