Pichai Admits Mistake
Google’s AI, ‘Gemini’ (formerly called ‘Bard’)recently found itself in an intense controversy following its problematic text and image responses. Google CEO Sundar Pichai has since responded, admitting that the company made a mistake. What happened was “completely unacceptable” expressed Pichai and that his company “got it wrong”.
“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,” Pichai wrote in a memo that was shared by Semafor.
The chatbot was deactivated the previous week following a social media uproar caused by its peculiarly revisionist images in what was described as an “absurdly woke” response.
What’s wrong with Gemini?
Over the past week, Google’s AI chatbot Gemini has been malfunctioning, erring and sinking deeper into controversy. It began when the chatbot generated inaccurate historical images, such as depicting the Pope as a woman and Vikings as Black people.
The situation exacerbated when Gemini failed to provide a definitive answer on whether Elon Musk posting memes or Adolf Hitler is worse, adding fuel to the controversy.
In a memo to the company’s employees, Pichai acknowledged the recent LLM’s (Large Language Model) glitches in text and image responses.
Nevertheless, he respectfully expressed awareness that certain responses from the app have offended users and displayed bias, stating that such outcomes are entirely ‘unacceptable’, and the company admits its mistake.
Does this make Google’s AI questionnable?
Occuring shortly after the rebranding of its chatbot and the introduction of the image generation tool just a few days earlier, this fiasco is a dent on the public’s trust in Google’s AI tools at a crucial juncture for the company.
Being aware of AI’s limitations, Pichai diplomatically pointed out in defence of Gemini, saying “No AI is perfect, especially at this emerging stage of the industry’s development.
In spite of saying this, Pichai maintained his firm’s goodwill and assure that the company will keep striving to meet the high bar set for them however long it takes.
‘We’re fixing the issues’ Assures Pichai
“And we’ll review what happened and make sure we fix it at scale,” an excerpt from the memo reads.
Pichai outlined several actions that will be taken to address the issue, including restructuring within the company, updating product guidelines, improving launchings, conducting more thorough testing, and recommending changes to their technology.
“We are looking across all of this and will make the necessary changes,” Pichai added.
He reiterated the company’s mission to organise the world’s information and ensure its universal accessibility and usefulness, considering it sacrosanct.
He stressed the importance of providing users with helpful, accurate, and unbiased information in all products, including emerging AI products, to maintain trust.
Gemini needs more fine-tuning
Despite Pichai’s defence, the issue lingers not just in the inaccuracy of the model but in the fact that Google prematurely rolled out its LLM.
There is a recognized need for additional fine-tuning of the Model. Google should conduct more comprehensive testing of its products to ensure they meet the necessary standards before being made available to the public.
Comments 1