Google explains what went wrong with Gemini AI image generation

New Delhi: As world leaders and industry stalwarts slammed Google over inaccuracies in its AI-generated historical images, the tech giant has tried to explain what exactly went wrong with its AI.

The company has made the decision to pause Gemini’s image generation of people while it works on “improving the accuracy of its responses”.

While Union Minister of State for Electronics and IT, Rajeev Chandrasekhar, expressed concern over the potential violation of Indian IT laws by Google’s Gemini AI chatbot, Tesla and SpaceX CEO Elon Musk accused Google of running “racist, anti-civilisational programming” with its AI models.

Prabhakar Raghavan, Senior Vice President at Google, admitted in a latest blog post that it is clear that “this feature missed the mark”.

“Some of the images generated are inaccurate or even offensive. We’re grateful for users’ feedback and are sorry the feature didn’t work well,” Raghavan said.

So what went wrong?

“In short, two things. First, our tuning to ensure that Gemini showed a range of people failed to account for cases that should clearly not show a range,” explained Raghavan.

And second, over time, the model became way more cautious than “we intended and refused to answer certain prompts entirely — wrongly interpreting some very anodyne prompts as sensitive”.

The company said that it did not want it to create inaccurate historical — or any other — images.

“So we turned the image generation of people off and will work to improve it significantly before turning it back on. This process will include extensive testing,” said Raghavan.

However, he said that he can’t “promise that Gemini won’t occasionally generate embarrassing, inaccurate or offensive results”.

“But I can promise that we will continue to take action whenever we identify an issue,” Raghavan added.
IANS

realme expanded its product strategy with Watch S2 to create AI-enabled ecosystem

New Delhi: The landscape of wearable technology is undergoing a transformative shift, fueled by the rise of artificial intelligence (AI). While wearables initially gained popularity for features like fitness tracking...

Google to empower 10,000 Indian startups in AI, unveils new tools

Bengaluru: Google on Wednesday said it is working with MeitY ‘Startup Hub’ to train 10,000 startups in artificial intelligence (AI), as the tech giant expanded access to its AI models...

Meta rolls out ‘verified subscription’ plans for businesses in India

Mumbai: Meta on Wednesday introduced verified subscription plans for businesses on Facebook and Instagram in India. The Meta Verified for businesses will offer them with a verified badge, enhanced account...

Tech company OnePlus unveils first 5G metal unibody smartphone & more

Milan: Global technology brand OnePlus on Tuesday unveiled four new products at an event here that includes Nord 4 which is the only 5G metal unibody smartphone ever. The company...

Apple approves first PC emulator for iOS to run classic games

New Delhi: Tech giant Apple has approved the first PC emulator for the iOS ecosystem to run classic software and games. The iPhone maker approved ‘UTM SE’ after initially rejecting...

Human-centric, responsible mobile AI is the future: Experts

Paris: Human-centric mobile artificial intelligence (AI) is the future and open collaboration among the industry and a distinct approach to hybrid AI is the only way to achieve the best...

Nothing sells 1 lakh units of India-made CMF Phone 1 in 3 hours

New Delhi: London-based technology company Nothing on Friday announced that 1 lakh units of its sub-brand's debut smartphone CMF Phone 1, which is being manufactured in India, was sold in...

Musk’s X developing ‘dislike’ button for downvoting replies

New Delhi: Elon Musk-run X is developing a downvoting feature which will be used to improve the ranking of replies.   While the company hasn't announced its plans officially, recent...

Indian scientists make breakthrough in quantum-based data encryption

Bengaluru: In a major breakthrough, a team of Indian scientists has created a user-friendly way to generate unpredictable random numbers crucial for stronger data encryption and provide robust cybersecurity, the...

India set to make strides in quantum technology, AI with the US: Union Minister

New Delhi:India, under the leadership of Prime Minister Narendra Modi, is set to make great strides in quantum technology, artificial intelligence, critical metals and semiconductor sectors with the United States,...

Five automakers to recall over 1,56,000 cars for faulty parts

Seoul: Kia, Nissan Korea and three other carmakers will voluntarily recall more than 1,56,000 vehicles due to faulty components, the transport ministry here said on Thursday. The five companies, also...

AI Chatbots Alexa, MyAI, Bing come with ‘empathy gap’, may harm children

New Delhi: Artificial intelligence (AI) chatbots like Amazon’s AI voice assistant Alexa, Snapchat’s My AI, and Microsoft’s Bing have frequently shown signs of an “empathy gap” that puts young users...

Read Previous

4th Test: Root leads fightback with unbeaten century to carry England past 300

Read Next

Gujarat: AAP, Cong reach seat-sharing deal; Ahmed Patel’s son misses out on Bharuch

Leave a Reply

Your email address will not be published.

WP2Social Auto Publish Powered By : XYZScripts.com