AI revolution or ethical dilemma? What 4.2 million tweets reveal about public perception of ChatGPT | usagoldmines.com

An evaluation of hundreds of thousands of English-language tweets discussing ChatGPT within the first three months after its launch revealed that whereas most people expressed pleasure over this highly effective new software, there was additionally concern about its potential for misuse. Damaging opinions raised questions on its credibility, potential biases, moral points, and issues associated to the employment rights of knowledge annotators and programmers. However, optimistic views highlighted pleasure about its potential use in numerous fields. The paper was revealed in PLOS ONE.

ChatGPT is a complicated AI language mannequin developed by OpenAI, designed to grasp and generate human-like textual content based mostly on consumer enter. It was launched to the general public in November 2022 as a part of the GPT-3.5 structure and later enhanced with variations like GPT-4. ChatGPT can carry out numerous duties, equivalent to answering questions, offering explanations, producing textual content, providing recommendation, and aiding with problem-solving. It makes use of deep studying methods to foretell essentially the most related responses, enabling it to interact in interactive conversations on a variety of subjects. The mannequin was educated on massive datasets, together with books, articles, and on-line content material, which permit it to generate coherent and contextually acceptable responses.

Whereas helpful in lots of areas, ChatGPT has limitations, equivalent to often offering inaccurate or biased info or producing fully fabricated responses (referred to as AI hallucinations). It has been utilized in numerous fields, equivalent to schooling, customer support, and content material creation.

When ChatGPT was first launched, its reputation soared, with its consumer base reaching 100 million people within the first month. Since then, many new AI language fashions have been developed by numerous corporations. Nonetheless, it may very well be argued that ChatGPT sparked the AI revolution within the office, producing widespread discussions about AI and prompting folks to kind assorted opinions about its affect.

Researchers Reuben Ng and Ting Yu Joanne Chow aimed to investigate the passion and feelings surrounding the preliminary public perceptions of ChatGPT. They examined a dataset containing 4.2 million tweets that talked about ChatGPT as a key phrase, revealed between December 1, 2022, and March 1, 2023—primarily, the primary three months after ChatGPT’s launch. The researchers sought to establish the problems and themes most steadily mentioned and essentially the most generally used key phrases and sentiments in tweets about ChatGPT.

The research analyzed the dataset in two methods. First, the researchers centered on figuring out vital spikes in Twitter exercise, or intervals when the variety of tweets, replies, and retweets about ChatGPT was notably excessive, they usually analyzed what customers have been saying throughout these occasions. They collected and analyzed the highest 100 most-engaged tweets from these intervals. Second, they recognized the highest key phrases every week that expressed optimistic, impartial, or damaging sentiments about ChatGPT.

The outcomes confirmed that there have been 23 peaks in Twitter exercise throughout the research interval. The primary peak occurred when ChatGPT surpassed 5 million customers simply 5 days after its launch, reflecting each the preliminary buzz and hesitancy surrounding the brand new software. The second peak was primarily centered on discussions about ChatGPT’s potential makes use of. Subsequent peaks explored its utility in educational settings, detection of bias, philosophical thought experiments, discussions of its ethical permissibility, and its function as a mirror to humanity.

The evaluation of key phrases revealed that essentially the most frequent damaging phrases expressed issues about ChatGPT’s credibility (e.g., hallucinated, loopy loop, cognitive dissonance, restricted data, easy errors, overconfidence, deceptive), implicit bias in generated responses (e.g., bias, deceptive, political bias, wing bias, spiritual bias), environmental ethics (e.g., fossil fuels), the employment rights of knowledge annotators (e.g., outsourced employees, investigation), and adjoining debates about whether or not utilizing a neural community educated on present human works is moral (e.g., stolen paintings, minimal effort).

Constructive and impartial key phrases expressed pleasure concerning the basic potentialities (e.g., big breakthrough, largest tech innovation), notably in coding (e.g., good debugging companion, insanely helpful, code), as a inventive software (e.g., content material creation superpower, copywriters), in schooling (e.g., lesson plans, essays, undergraduate paper, educational functions, grammar checker), and for private use (e.g., exercise plan, meal plan, calorie targets, customized assembly templates).

“General, sentiments and themes have been double-edged, expressing pleasure over this highly effective new software and wariness towards its potential for misuse,” the research authors concluded.

The research gives an attention-grabbing historic evaluation of public discourse about ChatGPT. Nonetheless, it’s price noting that the research centered solely on English-language tweets, whereas a lot of the broader dialogue occurred outdoors of Twitter and in non-English languages.

The paper, “Powerful tool or too powerful? Early public discourse about ChatGPT across 4 million tweets,” was authored by Reuben Ng and Ting Yu Joanne Chow.