How researchers will evolve with Chat GPT

Omar Al-Ubaydli
Omar Al-Ubaydli
Published: Updated:
Read Mode
100% Font Size
6 min read

The November 2022 release of artificial intelligence wizard ChatGPT has left many academics fearing for their livelihoods, both as scholars whose research skills are made redundant, and as teachers who can no longer prevent students from plagiarizing. While it is too early for university professors to go the way of carriage coach people, lift operators, and other jobs lost to technology, they will need to evolve to keep earning their paycheck.

For all the latest headlines, follow our Google News channel online or via the app.

Many people first heard about ChatGPT due to its ability to competently respond to entertaining questions. For example, when my colleague gave me access (you currently need a VPN to use it in Bahrain), the first question I asked it was: “Write a poem about electrical engineering in the style of Wilfred Owen,” and I was astounded by the excellence of the response.

Once the amusement and giggles subside, however, researchers like me quickly commence introspecting existentially: will I still be in a job in 10 years? Why would my boss ask me to prepare an intelligence brief that takes me two days when he can ask ChatGPT and get a better answer in literally five seconds?

The pain is particularly acute for PhD holders in natural and social sciences because for decades, they have been insulated from the threat that artificial intelligence poses to their livelihoods. Economists regularly study the changing labor market as robots and computers displace humans, safe in the knowledge that these entities were too dumb to ever make an economist redundant.

Our skills are too nuanced, we told ourselves, as we continued to dispassionately analyze the possibility of administrative staff, factory workers, and soldiers turning into something children read about in history books, rather than growing up aspiring to become.

But now the shoe is on the other foot, as the technically proficient literature reviews that used to be a key source of our added value transformed into something that can be executed with a keystroke.

Once my hyperventilation subsided, I began to think about what skills I can still bring to the table in the age of ChatGPT. My first thought is that like all artificial intelligence technologies, ChatGPT is based on proficient analysis of existing knowledge. That makes it excellent at synthesizing previous work since it can read and absorb much more quickly than any human.

Its AI foundations also makes it pretty good at predicting things based on what we already know. However, for the time being, one blind spot it has is in formulating research plans and gathering new data, handing humans the advantage there – for now at least.

To illustrate this, suppose I want to predict the impact of Neom on the Saudi economy. While I am yet to pose this question to ChatGPT, my guess is that it would yield a respectable answer based on a retrospective look at the effects of other megaprojects inside and outside the Kingdom. However, such a response would be inadequate for two reasons.

First, unlike the UK, US, and other western economies, the volume of existing research-based knowledge about the Saudi economy is limited, even more so if we restrict our attention to indexed research. There are thousands of papers analyzing the effect of the Industrial Revolution in the UK, but only dozens looking at the effect of Saudi megaprojects.

Second, Neom is unique and unprecedented in many ways, and the world today differs from the world 10 years ago by a lot, limiting the value of past experience in predicting the future impact. Taken together, these two factors clearly mean that we need to gather and analyze new data, preferably on a continuous basis, if we want to acquire a refined understanding of how Neom will affect the Saudi economy.

ChatGPT can probably make suggestions about the data to gather, but it will be a poor substitute for a researcher who has access to the board members and other key personnel. One could imagine these people being interviewed by ChatGPT, but they would probably be more willing to give their time to a physical human who knows how to laugh, reassure, sympathize, and so on spontaneously and earnestly.

For physical phenomena that need to be measured, one can imagine an updated version of ChatGPT commanding drones to go out and gather data, such as humidity, land area, traffic patterns, and so on. But the highest quality modern research is based on an effective combination of quantitative and qualitative data, with the latter often emerging from face-to-face interviews with stakeholders.

Therefore, for the time being, I imagine that ChatGPT will gobble up a significant volume of researchers’ market share in conducting literature reviews and other retrospective syntheses of what we already know. That will push scholars into the areas where they still possess an advantage over computers: innovative research designs that emphasize gathering qualitative data from humans face-to-face.

The clock is ticking, however, and it’s not impossible to imagine even that niche being bulldozed by ChatGPT’s fifth or sixth incarnations, as androids that humans can’t even identify as being artificial befriend them and extract data from them according to sophisticated research protocols. But at least for the next decade or so, expect ethnographic methods to undergo a renaissance as PhD holders scramble to maintain relevance.

Read more:

New US sanctions on Iran’s IRGC Cooperative Foundation, senior officials

Doomsday Clock 2023: Will humanity move closer to apocalypse amid nuclear war tension

Disclaimer: Views expressed by writers in this section are their own and do not reflect Al Arabiya English's point-of-view.
Top Content Trending