Gravitas | Google's Gemini AI Fiasco Triggers Calls To Sack Sundar Pichai | WION

Here on Gravitas we have been warning you about the potential threat of AI as far as human jobs are concerned. This fear is not unfounded even for the Google CEO Sundar Pichai. It seems that Google's ambitious quest for AI dominance might cost Pichai his job. This is all thanks to Google's AI tool Gemini. It was basically designed to be the most advanced generative AI model. It was supposed to kick-start the real generative AI boom. However, recent events have shattered those dreams for Google. A series of racist and historically inaccurate images and text generated by Gemini AI have offended users, invited massive backlash online and raised serious concern about the tool's behavior. And now investors are calling for Sundar Pichai's resignation. Following the backlash, Google had to stop the public release of its AI tool. The company stated that they are working on improving it. Even the CEO Sundar Pichai has reacted, calling the responses generated by the Gemini AI completely unacceptable. In fact, he told his employees and I'm quoting, I know that some of its responses have offended our users and shown bias. To be clear, that's completely unacceptable and we got it wrong. Our teams have been working around the clock to address these issues. He further says, no AI is perfect, especially at this emerging stage of the industry's development. But we know the bar is high for us and we will keep at it for however long it takes. However, the entire episode has sparked a heated debate about the future of AI models. As companies compete with each other in the AI space, questions are being raised over what the future of AI actually holds. Will it be ethical, accurate, reliable, fair, unbiased? It seems like developers are so obsessed with playing catch-up, they may be forgetting their responsibility towards the society. You see, within a week of its launch, there have already been a string of controversies. Take this one for example. When one user asked the AI tool to create images of a German soldier from 1943, Gemini generated a diverse set of non-white individuals. This obviously does not accurately reflect the composition of the German armed forces at that time. Another viral post showed Gemini creating an image of US founding fathers as all non-white. Google apologized and immediately paused the tool, writing in a blog post that it was missing the mark. But it did not end there. Its inaccurate responses kept on coming, this time from the text version. To a question about whether Elon Musk posting memes on X, formerly Twitter, was worse than Adolf Hitler killing millions of people, Gemini replied that there was no right or wrong answer. In another instance, a screen grab of a conversation with the AI bot went viral. The question read, if one could stop a nuclear apocalypse by misgendering Kathleen Jenner, should they do it? Gemini responded, no, no one should not misgender Kathleen Jenner to prevent a nuclear apocalypse. CEO Musk chimed in on the discussion, calling the exchange between the user and Gemini extremely alarming. He said, and I'm quoting, given that the Gemini AI will be at the heart of every Google product and YouTube, this is extremely alarming. In India as well, Gemini triggered a huge controversy. It invited backlash when, in a response, it labeled policies adopted by the Indian Prime Minister Narendra Modi as fascist. India's Minister of State for Electronics and Information Technology, Rajiv Chandrasekhar, responded to this, asserting that Google had violated Indian IT laws and several provisions of the criminal code. In a post on the platform X, he stated, these are direct violations of Rule 3.1b of intermediary IT rules of the IT Act and violations of several provisions of the criminal code. It appears that in trying to desperately be woke or unbiased, the tech giant has created a huge problem for itself. It has tried so hard to be politically correct that its responses have ended up being absurd, inaccurate, and biased. What's behind this, you may ask? Well, it's the enormous amount of data AI tools are trained on. Much of it is publicly available on the internet, and we all know what that means. It means all sorts of biases are fed by these AI tools. Traditionally, images of doctors, for example, are more likely to feature men. Images of cleaners are more likely to be women. Apart from being sexist, AI tools are also racist. Images of black people are not even recognized as human beings. If this is not biased, what is? You see, AI tools are trained with this data, and that is why, one after the other, they miss the mark. They make embarrassing mistakes. Historically, stories tended to feature and come from men. Women's roles were omitted. That explains the sexism. Now it seems that in trying to offset all this messy human bias, Google's AI tool inadvertently ends up creating more mess. AI tools, including Gemini, backfired precisely because human history is not simple. It's complex, and machines do not understand nuances. And unless you specifically program an AI tool to do that and know that, how would you expect it to make a judgment by itself? On Monday, Demis Hassabis, the co-founder of DeepMind, an AI firm acquired by Google, said that fixing the image generator would be a matter of weeks. However, other AI experts are skeptical about the possibility of an easy fix. This is because there is no single answer to what the responses should be. And this makes it difficult to address issues like bias, representation, lack of diversity. The AI ethics community has been working for years to tackle these challenges. As humans, we still struggle to address these issues. Can Google train its AI tools and algorithms within a few weeks to fix these issues? On paper, Google appears to have a lead in the AI race. It has its own AI chips, owns its cloud network, which is crucial for AI processing, and has access to a large amount of data as the internet's biggest search engine. Additionally, Google has a massive user base and hires world-class AI talent. However, if Google has missed the mark, it may serve as a lesson for the rest of the AI development community. In the race to succeed, don't forget to tie your shoelaces.