Critical thinking is a must for mastering AI

October 14, 2024 22:24

Since OpenAI's ChatGPT emerged at the end of 2022, a variety of generative artificial intelligence (AI) software has emerged.

Many academic journals are struggling on whether the use of AI should be allowed in writing research papers, and even if it does, it should be restricted to a certain percentage. For example, the American Association for the Advancement of Science (AAAS), which has a history of more than 140 years and includes well-known journals such as Science, recently gathered 4,000 researchers from different disciplines and countries to “weigh in on guidelines that could be adopted widely across academic publishing”. It is estimated that tools such as ChatGPT may be approved to contribute to 1% to 5% of manuscripts. The Journal of the American Medical Association (JAMA), of which the acceptance rate of the more than 9,500 annual submissions is only 12%, requires authors to “take responsibility for the integrity of the content” generated by AI related tools.

However, in the creative world, including article writing, art design and music composition, there have been many suspected AI productions. Because there is no need to declare the proportion of AI content as in academic papers, many people want to distinguish between AI and non-AI works. A number of articles on the internet have used this as the theme. Some say that AI-written articles are dull and repetitive, or overly dramatic with awkward sentence structures and phrasing. Some cite "common AI words", such as “however”, “facilitate”, “innovative”, “efficiency”, “in conclusion”... However, these are also words commonly used in our daily life!

Some attempt to use AI detectors to indicate AI-written text and find it very challenging. Even the AI classifier produced by OpenAI, the accuracy is less than 30%, no wonder the company admits that the tool “is not fully reliable”.

Some point out that if you use prompt engineering skillfully and guide AI as needed, you can create unique works. For example, you can first ask AI to write a first draft of a novel with the writing style of J.K. Rowling (author of the Harry Potter series), then tweak some passages with shock, romance or mystery elements, just like a director guiding the actors and film crew, and tailors the style to make the production a masterpiece that looks like works by a real human being.

Currently, we are still in the early stages of AI adoption, and over time, our skills will become more mature. However, as the world embraces these advanced technologies, we also need to think from all perspectives.

Yuval Noah Harari, an Israeli historian whose books have sold over 45 million copies in 65 languages, and is considered one of the world’s most influential public intellectuals, warned earlier on social media that the potential of AI to make decisions on its own is worrying. Nobody knows where AI will lead humanity to, and the epoch-making changes it will bring. He cites the Industrial Revolution driven by the invention of steam engine as an example. Today it seems that the steam engine has indeed improved human life, but the great leap forward in productivity has also triggered a huge demand for raw materials and markets, which has led to continuing global human conflicts over the past 200 years.

Though the argument sounds negative, it is also a good reminder. Therefore, while everyone is chasing AI, don't forget to uphold our critical thinking which is the crucial way that humans can take control of their own future.

Adjunct Professor, School of Computing and Data Science;Department of Geography, Faculty of Social Sciences, The University of Hong Kong

Most Popular 24 Hrs