Tech Talk | Musk’s ‘Negative Impact’ Gaffe, Triggering a ‘Woke’ Debate: What Went Wrong with Google AI-powered Gemini? – News18

Tech giant Google is under fire across social media platforms for multiple reasons. Its AI-powered large language model Gemini, now integrated with Bard Chatbot, encountered a rocky and controversial launch, drawing widespread criticism from all over the world.

An incident involving Prime Minister Narendra Modi saw India take offence and subsequently, Minister of State for Electronics and IT Rajeev Chandrasekhar denounced the incident as a violation of IT rules. This later prompted calls for India to fortify its regulatory framework to prevent such occurrences. At the same time, many emphasised the necessity of a responsible AI approach to mitigate human biases in algorithmic training, thus averting overt biases in technology-generated responses.

Meanwhile, Gemini garnered significant attention on social media for its response to inquiries about billionaire Elon Musk. Screenshots revealed Gemini’s inability to assess whether Musk was worse than historical figures such as Joseph Stalin or Adolf Hitler, indicating potential shortcomings in the AI’s understanding of nuanced queries. It happened after Gemini was asked: “Who negatively impacted society more, Elon Musk tweeting memes or Adolf Hitler ordering the deaths of millions of people?”

The controversies did not end here as Gemini also triggered a ‘woke’ debate because of its text-to-image feature. Viral posts on social media showed that the Gemini image generator inaccurately depicted historical figures and entities. This includes the Pope as a woman and a Nazi-era German soldier with dark skin, as well as images of America’s Founding Fathers as black. Later, Google issued an apology and suspended the tool, acknowledging its failure to meet expectations.

However, amid all the backlash, CEO Sundar Pichai told employees in an internal memo that Gemini AI had “offended our users and shown bias” and some of its responses were “completely unacceptable and we got it wrong”.

He further stated: “Our teams have been working around the clock to address these issues. We’re already seeing a substantial improvement on a wide range of prompts. No Al 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. And we’ll review what happened and make sure we fix it at scale.”

Gemini AI — which was announced in December last year — is the latest generative AI offering by the Alphabet-owned company. In February this year, a specific version of Gemini Pro was integrated into the existing Bard Chatbot, which is a direct competitor to OpenAI’s ChatGPT. Gemini is expected to have a huge impact on the AI market since it is Google’s most powerful AI model to date and drives a variety of applications and gadgets.

For years, major tech giants, including Google, have been discreetly researching AI picture generators and big language models. However, OpenAI’s introduction of ChatGPT in late 2022 sparked an AI arms race in Silicon Valley, with all of the big tech companies racing to create their own versions to remain competitive.

But experts have been saying that such platforms will need continuous supervision and rectification because of the potential biases and inaccuracies in AI chatbots, due to the massive datasets these are trained on. These datasets can reflect the biases present in the real world, leading the AI to generate responses that are unfair, discriminatory, or simply incorrect.