After Gemini, India's Krutrim AI chatbot faces flak for inaccuracy
What's the story
Krutrim, an AI start-up created by Ola founder Bhavish Aggarwal, unveiled an AI chatbot for public beta testing on February 26.
However, users have pointed out that the chatbot seems undertrained, providing inaccurate results in areas like general queries, translation, math problems, and logical reasoning.
Similar to OpenAI's ChatGPT and Google's Gemini, this chatbot uses multi-lingual, large language models.
Inaccurate
Factually incorrect answers and struggle with basic questions
During the testing phase, Krutrim's chatbot made several mistakes, such as incorrectly stating that the West Indies won the 1983 Cricket World Cup instead of India led by Kapil Dev.
It falsely claimed that Hillary Clinton defeated Donald Trump to become US President in 2014.
Additionally, the chatbot struggled with basic questions like today's date, who formed the Indian government in 2014, who killed Mahatma Gandhi, and who was India's first female Prime Minister.
Questioned
Questions raised about Krutrim's model and investment
These blunders have raised concerns about Krutrim AI's model and the substantial investment in the start-up.
Krutrim became India's first unicorn in 2024 after securing $50 million from Matrix Partners India, a former supporter of Aggarwal's ventures Ola Cabs and Ola Electric.
When asked about India's first billion-dollar start-up, the bot wrongly stated that its parent company Krutrim was the first Indian unicorn, while InMobi was actually the country's first billion-dollar start-up.
Larger issue
AI bots facing issues worldwide
With the advent of artificial intelligence, even major chatbots by giants like Google are facing flak globally.
Google boss Sundar Pichai recently apologized for Gemini's embarrassingly wrong answers.
Akin to Krutrim, Gemini provided incorrect answers to questions like America's founding fathers and Google's own co-founders.
Google has suspended its image generator AI tool for generating controversial images like the ethnically diverse Nazi German army during WWII.