Skip to main content

Users of ChatGPT will be able to programme the viral chatbot to reflect their own beliefs, raising the prospect of rival Left and Right-wing versions being created.

OpenAI, the organisation behind ChatGPT, said it was developing an upgrade that would let users more easily customise the artificial intelligence system.

It comes after criticism that ChatGPT exhibits a Left-wing bias when answering questions about Donald Trump and gender identity. The bot has described the former US president as “divisive and misleading” and refused to write a poem praising him, despite obliging when asked to create one about Joe Biden.

ChatGPT has also said it is programmed not to promote fossil fuel use and has said it is not able to define a woman, raising criticism from US Republicans.

OpenAI said: “Since our launch of ChatGPT, users have shared outputs that they consider politically biased, offensive, or otherwise objectionable. In many cases, we think that the concerns raised have been valid and have uncovered real limitations of our systems which we want to address.”

The company said that it takes steps not to favour political viewpoints when developing the software, but that biases inherent in the billions of sentences it is trained on may be reflected by the software.

However, it said that it wanted individual users to be able to customise the bot, even when it meant that people would disagree with it.

The company added: “We believe that AI should be a useful tool for individual people, and thus customisable by each user up to limits defined by society. Therefore, we are developing an upgrade to ChatGPT to allow users to easily customise its behaviour.

“This will mean allowing system outputs that other people (ourselves included) may strongly disagree with.” 

OpenAI said it would continue to impose some limits on different versions of ChatGPT. “Striking the right balance here will be challenging – taking customisation to the extreme would risk enabling malicious uses of our technology and sycophantic AIs that mindlessly amplify people’s existing beliefs,” it said.

It added that different versions of the bot will tackle the risk of the company having a “single, monolithic AI system”.

ChatGPT, which allows users to ask factual questions, pen essays, songs and poems, and help write code, has become a near-instant sensation since its launch in November, with a reported 100m users.

Microsoft’s Bing search engine has integrated technology similar to the bot, although early testers have this week reported bizarre conversations involving it turning hostile and seeking to break up users’ marriages.

OpenAI released a portion of its guidelines for the chatbot. They show that ChatGPT is prohibited from inciting hatred or violence, influencing politics or writing adult material.


About SDX

Leave a Reply

%d bloggers like this: