“Google Issues Apology to Government Over Gemini’s Inconsistent Results on PM Modi”

NEW DELHI: American tech giant Google issued an apology to the government for the unsubstantiated comments made by its AI platform Gemini regarding PM Modi, acknowledging its platform as ‘unreliable,’ Minister of State for IT & Electronics Rajeev Chandrasekhar informed TOI.

“We had sent them a notice, seeking an explanation on the unsubstantiated results thrown up by Gemini regarding a particular query on PM Modi. They replied, stating, ‘Sorry, the platform is unreliable,'” the minister revealed, coinciding with the government’s announcement that AI platforms will now require a ‘permit’ to operate in the country. “That’s an inadequate response,” criticized the minister, highlighting a concerning trend among AI platforms offering ‘consumer solutions’ during trial phases.

The minister emphasized that India cannot serve as a testing ground for AI platforms, particularly as they face increasing global scrutiny for providing biased, misinformed, or unverified results. “AI data is being released from labs straight onto the public internet, without proper testing or safeguards. Then, when caught, they apologize, citing unreliability.”

Citing Google’s Gemini as an example, he stated, “Google Gemini is a prime example. They transitioned from the lab to the public internet without fear of consequences. And when confronted, they dismiss it, claiming the information is unreliable.”

The minister underscored that the Indian government will not permit underdeveloped platforms to offer full-fledged services without proper disclosures to users regarding potential misinformation or unlawfulness.
“We’ve made it clear that the Indian internet is not a testing ground. If you’re transitioning from the lab and it’s still unreliable, you must include a disclaimer. Users must be explicitly informed through consent forms and terms of use that this platform is under trial and may produce inaccurate or unlawful content.”

The minister emphasized that AI platforms cannot treat India as an extension of their research and development laboratories. “Indian internet and consumers cannot be considered mere testing grounds. Respect for Indian consumers, our digital citizens, is paramount. It is imperative for AI platforms to explicitly warn Indian users about potential errors, unreliability, and inaccuracies in their platforms. Users must be fully aware and informed before engaging with such platforms.”

Furthermore, the minister highlighted the legal repercussions for AI platforms found violating Indian IT and criminal laws.
“The Indian legal framework, including the IT Act and IT Rules, clearly delineates 12 types of unlawful content. If AI platforms generate such unlawful content, they cannot claim unreliability as a defense. Prosecution will ensue.”

The minister condemned Gemini’s dissemination of unverified and unsubstantiated results as “reckless, irresponsible, and disrespectful” towards Indian consumers. “Taking Indian digital citizens for granted will not be tolerated, and appropriate measures will be taken to address such behavior.”

The government’s recent advisory for AI-led startups underscores the importance of labeling unverified information as potentially false and error-prone. This advisory follows a previous notice issued nearly two-and-a-half months ago regarding deepfakes, in response to numerous instances of synthetically-generated content infiltrating social media and online platforms.