If Chatgpt was writing this article, probably it will start with an introduction to AI and magical capabilities of it.
However, as an ordinary human being, I am typing this article on a flight to a new destination while thinking about how our world has evolved and will continue to evolve with the exponential tech developments. (just to give you an idea, by the time Zuckerberg woke up on Thursday, July 6th, the number of downloads of Thread app jumped from 10M to 30M)
But despite the rapid inventions, like all other ordinary humans, we all have our own likes, dislikes, belief systems, and biases. In fact, bias is a part of our human mind’s defense mechanism. Simply to keep us alive by maintaining our well-known, predictable environments, places, rules, and systems to avoid any danger (yes you guess right, it is aka “comfort zone”)Whether we are aware or not aware, we all bring our own perspectives to conversations and relationships, to schools and to workplaces.
Then the question is what about AI biases? In the end, we all know that AI has been designed by humans and where are those humans’ perspectives located in AI algorithms?
To put it simply, AI bias occurs because human beings choose the data that algorithms use, and also decide how the results of those algorithms will be applied. Without extensive testing and diverse teams, it is easy for unconscious biases to leak machine learning models. Then AI systems automate and perpetuate those biased models over and over again.
I couldn’t help but ask the same question to Chatgpt! (Here is the prompt I use for curious minds)
“From 1 to 10, how do you rate AI for having biases? 1 no biases, 10 very biased”
Here is the response I got:
“As an AI language model, I don’t have personal opinions or experiences. However, I can provide you with some information.
The extent to which AI systems exhibit biases can vary depending on several factors.
These factors include the data used to train the AI, the algorithms and models employed, and the design choices made during development. Bias can arise in AI systems when the training data contains biased information or when the algorithms themselves inadvertently introduce bias.
On a scale of 1 to 10, it is difficult to give a universal rating for AI biases because it depends on the specific AI system and its implementation. Some AI systems may exhibit minimal biases(closer to 1), while others may demonstrate more noticeable biases (closer to 10).
It’s important to note that efforts are being made to reduce biases in AI systems and ensure fairness. Researchers and developers are working on techniques such as de-biasing algorithms, diverse training data, and rigorous evaluation processes to mitigate biases in AI. However, complete elimination of biases is a complex challenge that requires ongoing attention and improvement.”
So my other question to my peers, tech companies and us as responsible global citizens:
What we start doing today to remove AI bias before it creates a perpetual fake reality and transform our perception about the world we are in?
