Rewriting History!

This week saw something really irritating (to me at least, hence the rant!). Google launched their lastest chatbot, Gemini, to much fanfare and then very quickly pulled it back down. Why? Because it’s woke. The lady above was generated when trying to get an image of a strong Norse viking man. Apparently, the Pope was similarly represented (a lady of colour) and on it went. This issue is dedicated to the question of how do we prevent human biases and “wokeness” from taking over. Let’s dive in…

As I said, Google was compelled to deactivate the image generation capabilities of its latest AI model, Gemini, last week following complaints that it defaulted to depicting women and people of colour when asked to create images of historical figures traditionally portrayed as white and male, including Vikings, Popes, and German soldiers. The company issued a public apology and stated its commitment to improvement. Furthermore, Alphabet’s CEO, Sundar Pichai, dispatched a mea culpa memo to staff on Wednesday. “I am aware that some of its responses have caused offence to our users and demonstrated bias,” it reads. “To be clear, this is completely unacceptable, and we have erred.” It does seem incredibly easy to offend people these days. Or am I alone in thinking that? Sorry if that offends!

However, Google's critics have not been mollified. In recent days, conservative voices on social media have underscored text responses from Gemini that they claim exhibit a liberal bias. On Sunday, Elon Musk posted screenshots on X showing Gemini stating that it would be unacceptable to misgender Caitlyn Jenner even if this were the only means to avert nuclear war. “Google Gemini is supremely racist and sexist,” Musk wrote.

A source familiar with the situation suggests that the uproar reflects the evolving norms regarding what is deemed appropriate for AI models to produce. The company is working on projects aimed at mitigating the kinds of issues observed with Gemini in the future, the source adds. I’m sure it is!

Why did this happen? Perhaps simply because Google hurried the development of Gemini. The company is evidently struggling to find the right pace for releasing AI technologies. It once adopted a more cautious stance with its AI technology, deciding against releasing a powerful chatbot due to ethical concerns. After OpenAI’s ChatGPT became a global sensation, Google shifted gears. In its haste, quality control seems to have been compromised.

And yet, it is not alone. I had the most frustrating conversation with ChatGPT whilst researching this. I tried to create a fun picture of a balck woman nazi stormtrooper and it wouldn’t comply because it couldn’t depict anything that promoted hate or violence! I asked for a cartoon! With the world being so polarised, and extremism on the rise everywhere, it seems to me that history is more likely to be repeated if we ignore its lessons. AI should be a force for good by presenting the unvarnished truth without being programmed to not cause offence and reflect some of the crazy biases that seem to be everywhere. Rant over…