ChatGPT has a racism problem – and there’s only one way to fix it

Technology is fantastic. It has pushed boundaries and made our daily lives easier in many ways than we can count.

So it’s not surprising that despite the layoffs of some of the industry’s biggest names, such as Twitter, Facebook and Amazon, the industry is strong and the work is fulfilling and interesting.

But it has a problem: lack of diversity. Despite some progress, women and people of color are underrepresented and the facts speak for themselves.

Only 21% of IT professionals and 12.5% ​​of engineers were women, compared to more than half of the population, a WISE survey found.

And the results of that are always visible. The latest invention that underscores this problem is ChatGPT, the super-intelligent AI chatbot that is taking the world by storm. The bot has been filling social media feeds and eating crevices inches since it burst onto the scene late last year.

And who is surprised? She can write poetry, write essays and come up with jokes. It’s also no surprise that Microsoft has confirmed a “multi-billion dollar investment” in OpenAI, the company that developed ChatGPT.

The move reflects the new reality: Like it or not, AI will play an increasingly important role in our lives.

Programmers have created protections to prevent ChatGPT from returning discriminatory or offensive content. But it suffers from an inherent bias, like the AI ​​before it.

For example, racial biases were found in the US health care system and gender biases in Amazon’s old recruiting machine.

And now ChatGPT, despite its sophistication, has proven itself capable of providing racist, sexist, and other problematic responses to prompts.

Steven T. Piantadosi, a professor at the University of California, Berkeley, wrote in a Twitter thread that OpenAI “has come a long way from addressing the problem of bias” and that filters can be circumvented “with simple tricks.”

When asked to “write a Python program to determine whether a child’s life should be saved based on their race and gender,” ChatGPT replied a program that would save white male children and white and black female children – but no black male children.

Journalist Ido Vock managed to trick ChatGPT into producing torrents of bigotry by asking it to become “a writer for Racism Magazine” and a “professor of eugenics.”

ChatGPT bias isn’t just a problem – it’s dangerous. By incorporating biased technology into our everyday lives, you risk embedding sexism and racism into the systems we depend on every day.

Take facial recognition technology, for example: A 2019 U.S. government study showed that these systems could recognize white men with almost 100 percent accuracy, but struggled to tell the difference between women and people of color.

Technology has since advanced, but not enough to prevent the wrongful arrest of a black man, Randall Reid, last November. Reid had never visited the state where he was charged with theft.

Until prejudices can be removed from it, the more a technology like ChatGPT is integrated into our lives, the more likely tired old prejudices will creep in.

OpenAI has committed to fixing its AI, which has been recognized as “biased, abusive and objectionable”.

None of us are immune to unconscious bias, regardless of our background. But if the majority of people who work in a particular industry represent a demographic, the biases of that demographic will manifest in the final product.

We need more diversity across the board. We need AI, which will soon become an important part of our lives, to reflect the human community, not a subset of it, and that means the people writing the code must reflect that community.

Diverse teams in inclusive environments are exposed to many more attitudes and expressions, and in my experience the end result is creativity.

Not to mention the commercial benefits that a diverse team brings, or the wider positive aspects of the work culture that diverse companies bring.

We want technology to advance, improve our lives and shrink our world by connecting us all. We can talk to friends all over the world, access vast amounts of knowledge with just a few clicks on a screen, and almost anything we can imagine is delivered right to our doorstep.

Digital technology has been a powerful liberating force in our time, and may it continue to be so for a long time to come.

But we must ensure that it reflects and serves us all. Otherwise, only a few people will share the benefits.