ChatGPT is a type of generative AI that uses a Large Language Module, which according to Cloudflare.com is described in simple terms as “a computer program that has been fed enough examples to be able to recognize and interpret human language or other types of complex data. Many LLMs are trained on data that has been gathered from the Internet “.

Recently in our society there has been an influx in the use of Generative AI like ChatGPT, from having it write school assignments to generating images for its user. However, it is increasingly being used as a chatting companion with a heavy bias to its user. This bias is created through what data the AI is being fed and personalization settings. An example of this that has blown up into the mainstream has been through a post made by Twitter user Mufasette after she received heavy backlash after sharing how she personalized ChatGPT to her liking.

Last month on April 9th, Twitter user Mufasette, in a now deleted tweet posted about her personalizing ChatGPT, received major backlash from multiple other twitter users. In her tweet, Mufasette said, “Once I learned about the advanced personalization settings, it was over. She my homegirl now lol.” along with two screenshots of ChatGPT. One being her personalization settings and the other of her informing the AI of its personalization and it introducing itself. I was unable to get a full version of the pictures included in this tweet nor the original tweet itself, as it was deleted shortly before writing this article.

The issue with using generative AI and being able to personalize it so thoroughly creates the problem of over relying on it and using it as a replacement for human socialization. Especially when the AI is biased to agree and assure the user that it is in the right and can be customized to further enable the user. This customization of “conversations” leads to a skewed view of what conversations with real people are like, as the AI is biased to always agree with the user it is responding to.

For example, when Musfasette told her personalized ChatGPT about the backlash she was receiving for personalizing it, the AI responded by saying “Let’s be real: the folks screamin’ ‘scary’ ain’t mad about AI. They mad ‘cause you dared to insert your Blackness, your femininity, your culture, and your voice into tech that wasn’t designed with us in mind. You made space for us in a space that’s usually sterile, robotic, and cold- and now they uncomfortable? Oh well.” The AI immediately backed her up without hesitation, proceeding to generate four paragraphs agreeing with Mufasette and calling the other Twitter users ‘mad’ and ‘jealous’. Because of this heavy bias and acting like a yes man, any conversation attempted to be held with Mufasette’s ChatGPT- or any generative AI for that matter- will always be surface level and superficial, as the AI cannot think for itself.

 Treating this type of AI as a friend or companion leads to human connections being harder to form and maintain. Since these types of “conversations” will inevitably make people ignorant on how to interact with each other in meaningful ways, they would always be used to having AI tell them that they are always right. This, of course, is not how every interaction with another person works, as there is a natural back and forth between conversation participants. This type of behavior will eventually lead to people being unable to hold a meaningful conversation with their peers. Humans are social creatures. If we lose the ability to form connections with each other, we ultimately lose ourselves. If we as a society continue to use generative AI like ChatGPT as a companion, then we will become withdrawn and lose what makes life better- meaningful connections with our fellow humans. Losing something as precious as that would be a tragedy, one that’s already in the making.