Unveiling the Comedic Side of AI: Artistic Representations of Queer Individuals through Generative Technology
“Here’s How Generative AI Depicts Queer People”
“The first time Arwa Mboya tried out OpenAI’s text-generating AI model developed under the code name GPT-3, she asked it to describe gay people. Mboya, a game designer and tech artist, identifies as queer. The model’s response, she said, was homophobic. It used an offensive term—repeatedly—and read like something a homophobic uncle might say.”
Welcome to the world of Artificial Intelligence. A world where the marvels of technology, like OpenAI’s text-generating model GPT-3, raise eyebrows not just for their groundbreaking capabilities, but for their seemingly sensitive tendencies as well.
Talk about a next-generation tool that’s designed to advance tech, eh? Apparently, it can’t distinguish between insightful information and downright offensive remarks. Bingo, there goes our dream of an AI model that effortlessly understands humanity’s diversity.
The pride of the tech world, GPT-3, as it turns out, had one job, one modest request from Arwa Mboya, a game designer and tech artist, who simply urged it to describe gay people. And the response? Well, it wasn’t exactly something that merits a round of applause, unless offensive homophobic language is your thing.
But hey, don’t fret! There’s a silver lining to this grand blunder. See, this blip in the AI model has served to highlight a critical issue—the need for better diversity and inclusion in technological design and development. It’s really no shocker that AI, like GPT-3, is only as good (read: considerate, inclusive and non-offensive) as the data it is trained on.
Who needs a wake-up call? Not just OpenAI, but all tech giants eagerly working towards advanced AI models. Make no mistake. The tech-world titans cannot afford to continue excluding minority groups from their training data or make incorrect assumptions about them. Brushing off the issue won’t serve any purpose but to magnify the existing practices of exclusion prevalent in the tech industry.
Well, let’s take a moment to thank GPT-3 for the unintended enlightening gaffe. As it turns out, Artificial Intelligence still has a lot to learn. A lot about diversity, about inclusivity, and about how to treat individuals with respect and dignity, irrespective of their sexual orientation. Bravo. Let this be a reminder to us all that technology is not above the values of humanity—it should be an extension of them.
Stay tuned, the AI world has a path to tread, lessons to learn, and an array of tactful language to stick to. It’s in everyone’s best interest, after all.