Does ChatGPT have a role in market research? 

OpenAI finished 2022 with an announcement to Google, Meta, and to the world that it is a serious contender in the AI chatbot race. No doubt the developers, engineers, and scientists at rival companies were working at full steam to release their own chatbots to the world. To the outside world, this is simply great technology and as Arthur Clarke wrote ‘any sufficiently advanced technology is indistinguishable from magic’. Except in this case, it is, and a quick examination will explain why. There are many definitions of what AI is as researchers have not yet reached a consensus so I will throw in my definition too ‘AI is an attempt to replicate human intelligence during decision-making or interpretation.’

The very idea of ChatGPT shows the capability of human intelligence. We built a plane and a few decades later we landed a man on the moon. As a species we have cured diseases, explored the oceans, transplanted hearts, and built self-driving cars. These were made possible by innovation, a human aspect even the most advanced AI systems would struggle with. I asked ChatGPT ‘Think of a world problem, and find the most effective way to solve it’. It selected climate change, no doubt a world problem, but its proposed solution can be summarised as making a transition to solar, wind and hydropower, incentivized by offering tax breaks/subsidies. This is not new technology nor a new incentive. Governments around the world have offered such incentives for years. I made a more specific request (garbage in, garbage out, right?) and this time explicitly asked it for a new way to solve a world problem. This time it chose poverty, and a summary of its proposed solution is  a basic income made to every citizen regardless of employment status, wealth, or other factors funded by a new global tax on wealth or carbon emissions. While the bot did acknowledge this would be a herculean task requiring significant international cooperation, political will and public support, it ultimately did not think of a new solution. The idea of a global basic income has been around for decades, and studies have been done on it. The bot does well to digest and summarise information on a topic – it deduced poverty is a global issue that is , but it falls short on critical analysis, judgement, and ability to conceive a new idea. This is why ChatGPT should be used to aide productivity, not to replace it.

Perhaps by nature of the industry, market researchers tend to be very creative with the use of new technologies. ChatGPT can serve as an arsenal in market research; from assisting with study designs to interpretation of results. In the project pipeline from concept creation to project debriefing, there is a role for ChatGPT to play. However, knowing and understanding your client is not something ChatGPT can do for you. The relationship between a client and a consultant can be unique, something not captured well by the bot. Another human strength that outmatches the bot is localized domain knowledge from previous experience on the project. ChatGPT is great for doing the more tedious/automatable tasks, but the true understanding of a project, client, and client needs is something developed over years of human experience.

Robin Mahachi - Senior Data Analyst