ChatGPT Could Produce More Homogenous Content

Few technologies have made as big a splash as ChatGPT in recent times. The generative AI tool was launched in November 2022 and quickly racked up a huge number of users as people began to use it to write everything from code to sonnets.

While the tool garnered a huge amount of headlines in those early months, with many afraid that it would result in a huge swathe of jobs being automated, the reality is that most of the content output by the tool still requires a decent amount of tweaking in order to fit our needs.

Finding the balance

As a result, you tend to need to strike a balance between speed and the quality of the output in meeting your needs and particular style. In other words, do we favor speed or style?

This trade-off can be pretty straightforward when we’re making it for ourselves, but when generative AI tools are deployed at scale, research from the Kellogg School suggests there are broader, more societally important consequences.

The researchers found that the main consequence of an increase in the use of generative AI is that the content that is ultimately produced will lose much of the novelty and ingenuity that is common for humanly-produced work and become more homogenous. What’s more, this will then become perpetuated as each wave of content is used to train the next generation of generative AI tools.

Retaining novelty

There are ways to lessen the problems AI can cause in society, say the researchers. One good way is to involve more people in using AI and letting them make changes. If the AI’s results match what users really want, we won’t have too much sameness or bias. But to make that happen, users need to tell the AI what they want.

In practice, this could mean the AI asking users some questions before answering to understand their style or point of view better. Or it might mean offering users a few different options to choose from.

For instance, the tools might give us a couple of versions and ask us to choose what we think is the best one. These suggestions will inevitably slow us down somewhat, and therefore potentially make the technology a bit less useful, but in the long-term, they will help both users and the tools themselves by ensuring more unique content is produced.

While many have voiced concern about the role generative AI will play in content creation, the researchers are more optimistic, but frame this optimism in cautious terms under the proviso that we have to retain the full range of human preferences.

This could be especially beneficial if the tools allow a new generation of people to be creators who might otherwise have not been, which could improve the diversity of content rather than reduce it. Time will tell which of these outcomes will ultimately materialize.

Facebooktwitterredditpinterestlinkedinmail