Artificial intelligence is everywhere these days, and everyone is afraid of losing their job. But researchers have now found that generative AI capabilities still fall short of replacing human scientists in most critical areas.
Large language models (LLMs) have already been incorporated into a variety of professional domains. The AI advancements are so rapid that many don’t even question whether the models will usurp their jobs – it’s almost a given.
However, at least in the field of academic research, it looks like AI still falls short of replacing human specialists, University of Florida researchers have found after asking generative AI models to write a research paper.
“We utilized commercially available AIs (e.g., ChatGPT 4) to perform each step of the research process, culminating in an AI-written manuscript. We attempted to intervene as little as possible in the AI-led idea generation, empirical testing, analysis, and reporting,” said the researchers.
In a paper titled “AI and the advent of the cyborg behavioral scientist,” they also detail how they limited any human intervention in all six stages of academic research.
What they discovered was a mixed bag of capabilities and limitations, and this is presumably good news for research scientists wondering if AI will take their job.
The paper, published in the Journal of Consumer Psychology, states: “We found that the AIs were adept at some parts of the process and wholly inadequate at others.”
Specifically, the researchers found AI to be a useful tool in the ideation process and research design, including the methods and stimulus design substages.
In the literature review, results analysis and manuscript production stages, though, it struggled to produce valuable outputs and required substantial oversight.
“A pervasive fear surrounding these AIs is their ability to usurp human labor,” explained Geoff Tomaino, an assistant professor in marketing at the University of Florida Warrington College of Business.
“In general, we found that these AIs can offer some assistance, but their value stops there, as assistance. These tools can do a great deal of legwork. However, the researcher still has a vital place in the process, acting as a director and critic of the AI, not an equal partner.”
The scientists’ recommendation is that behavioral researchers use AI models “judiciously” and carefully monitor the outputs for quality and coherence. Additionally, academic journals should consider adopting policies that call out AI assistance in research papers.
Earlier this year, Kristalina Georgieva, Managing Director at the International Monetary Fund, predicted that AI would impact 60% of jobs in advanced economies and 40% worldwide.
Even though job cuts haven’t so far been on the scale that many have predicted, experts think this could change if and when a recession strikes – and academic institutions aren’t immune to economic downturns.
Your email address will not be published. Required fields are markedmarked