Brilliant but Biased: Of Humans and AI
Now that I have AI, what is it that I can’t do?
I can paint, write, make videos, and even speak in another language. I can craft a speech in any tone of voice. With enough editing, I can sing flawlessly. A close friend once said to me, “You know, Tavleen, I think AI can complete our imperfections.”
I quickly asked if she thought she was imperfect. “We all are,” she replied. “Isn’t it amazing that some of our imperfections can be improved through technology?”
She was partially right. But how far does this really go?
Artificial intelligence is undeniably a breakthrough. It’s transforming the way we communicate and create. We use it constantly—whether we realize it or not. From facial recognition on our phones to automated customer service, from Siri and Alexa to self-driving cars and generative tools like ChatGPT, AI is everywhere.
The New Language of Imperfection
AI systems are trained on vast datasets designed to mimic human behavior and reasoning. But these datasets are far from perfect. Bias—often unintentional—frequently seeps into the data, and it shows in the results.
For example, a 2018 MIT study revealed that facial recognition systems, including the ones used to unlock smartphones, were significantly less accurate for people with darker skin tones—especially women.
Researchers at Stanford University have discovered that Large Language Models (AI) still exhibit decades old racial stereotypes in a covertly manner. AI in a manner is reflecting human biases too. But why do these biases happen?
One reason lies in the people behind the technology. Developers who create algorithms fill in biased data because the majority of the developers are from Western, Educated, Industrialized, Rich and Democratic societies, a term that experts call WEIRD. These backgrounds don’t represent the full diversity of the global population, and neither do the datasets used to train AI.
As a writer, I often rely on high-quality stock images for my work. Many AI-powered tools help generate these images. But when I search for visual content, I notice a clear Western bias. Asian representation is minimal. It’s not just an oversight—it’s a consequence of training data that doesn’t reflect the world in its entirety.
Use the Brilliant for Your Brilliance
Understanding how to use artificial intelligence—especially when it comes to young minds—is essential if we want to maximize its benefits while minimizing harm. A recent MIT experiment revealed that the groups that relied on generative AI to write essays became increasingly dependent on it, showing signs of reduced cognitive effort over time.
This raises an important question: Should children, whose brains are still developing, be allowed to rely heavily on AI? Well, I think not.
That said, this is just one part of the story. In other aspects, AI is undeniably remarkable. I use it to polish my writing, not to replace it. (In fact, it helped me fix grammar in this article.) My colleagues use it to streamline labor-intensive reports—but not to make important business decisions. Like any tool, its impact depends on how we choose to use it: to make us sharper—or simply more passive.
Flawed, and All the More Human
AI is an excellent tool to improve your methods, and amplify your strengths.
If you are not brilliant, neither is AI. It can generate content, but it cannot create connections. AI cannot laugh with a friend, rant with a sibling, play with your child, or hold your hand in a moment of pain.
Yes, it can help us manage our flaws—but it cannot erase them. And maybe it shouldn’t. Our flaws are what make us human afterall.
Tavleen Kour, soft skills trainer, and a writer.
