Chanel’s CEO, Leena Nair, recently shared an example of where ChatGPT, falls short. Visiting Microsoft’s Seattle headquarters, Nair tested ChatGPT with a straightforward prompt, only to find herself disappointed by the bot’s response that clashed with Chanel’s identity.
During her interview at the Stanford Graduate School of Business, Nair described asking ChatGPT to generate an image of Chanel’s senior leadership team visiting Microsoft. Instead of a diverse, inclusive representation, the AI produced an all-male team in suits hardly the image one would expect for a company that prides itself on its deep connection to women. “This is Chanel,” Nair pointed out, emphasizing that 76 percent of Chanel’s organization and 96 percent of its clients are women, a fact that should shape how AI interacts with or represents the brand.
For Nair, this incident reflects a broader issue with AI: its tendency to reflect the biases inherent in the data on which it’s trained. Despite claims of neutrality, AI systems frequently exhibit cultural and gender biases, a flaw exacerbated by insufficient guardrails against outdated or offensive depictions. While some have criticized AI for being “woke,” these technologies often err in the opposite direction, showing a lack of awareness or sensitivity to diversity.
Chanel is currently moving toward becoming “AI-ready,” but Nair stresses the importance of designing AI with ethics and integrity at the forefront. She notes that she’s consistently urging tech leaders to integrate “a humanistic way of thinking” into their AI designs so that they better represent the values of inclusivity and equality that many companies, including Chanel, stand for.
However, after Nair’s experience, it must be believed that the industry has a long way to go. AI systems remain heavily influenced by the biases and limitations of their training data, with generative models often revealing blind spots in their understanding of nuanced, culturally sensitive topics.
Nair’s story reflects that although AI is a powerful tool, its development must be guided by ethical considerations, especially as these systems become more embedded in our lives.