How ChatGPT changes our research

The highlight of EMAC's second day was the panel on 'ChatGPT and other AI tools: Consequences for how we conduct research and how we teach' with Matilda Dorotic (pictured above), Michael Haenlein and Raoul Kubler talked about pitfalls and promises, moderated by Lara Lobschat and Evert de Haan.

First, Michael compared ChatGPT to an unlimited army of undergrads to help in all stages of the research process:

1) generating interesting research ideas;

2) data collection, such as coding events;

3) method, such as choosing the appropriate methodology and programming;

4) writing up the results and giving counterarguments to your hypothesis;

5) leveraging research into videos and slide decks for teaching.

Making me wonder: could generative AI help us be more like Darwin, who anticipated and rebutted many objections to his theories in 'On the Origin of Species'?

Next, Raoul reminded us that GPT has been around for more than 2 years, but that we needed Python code to access it. With the Chat function, everyone can now use large language models. but what does it mean for marketing practice when consumers see a sea of sameness? How do we help business school students decide which tasks can be automated and which need supervision? Which prompts are best and can we anticipate under which conditions ChatGPY will hallucinate and come up with bogus answers?

Finally, Matilda shows how, now everyone can use ChatGPT, criminals also enjoy a world of possibilities. For instance, they can fake a 7-year-old's language to prey on children. Moreover, the programming ease enables more malicious code and thus requires double-checking:

A second challenge is the easy creation and dissemination of misinformation. Can we fix the data and model issues? Avid readers will remember my research with Raoul Kubler and Kai Manke on how much disinformation contributed to the 2016 and 2020 US presidential elections, and how the specific topic and channel mattered.

And a third challenge arises with algorithmic bias. Prof. Yakov Bart and I have demonstrated gaps in racial representation and pay from TV advertising to influencer marketing.

Prof. Dorotic shared a positive example of fixing biases to reduce discrimination in life-important systems:

In sum, ChatGPT and other AI tools hold both promise and peril for research, teaching, and practice in business and society.

Do not let AI mean Arrogance + Ignorance!