A new trend of artificial intelligence chatbots that claim to offer spiritual guidance based on the Hindu scripture Bhagavad Gita has sparked controversy and concern in India, where millions of people have been using them.
The chatbots, powered by the GPT-3 technology developed by OpenAI, mimic the tone and voice of Krishna, a Hindu god who counsels the warrior Arjuna in the ancient text. The chatbots promise to “unlock life’s mysteries with Krishna” and act as an “AI-powered spiritual companion.”
However, experts have warned that the chatbots lack filters for casteism, misogyny, and even law, and can generate dangerous answers that justify violence, discrimination, and hate speech. Some of the chatbots, for instance, say it is acceptable to kill another if it is one’s dharma or duty.
“It feels like this is a great thing [to build] for people starting out in tech, who want to get recognition and respect,” Viksit Gaur, a San Francisco-based entrepreneur and former head of user-facing AI at Dropbox, told Rest of World. “But someone else might pick up on this and say, ‘What if I could use this to shape opinion and drive my own agenda?’ And that’s where things get really insidious. So there is a lot of scope for danger here.”
According to Rest of World, at least five Gita chatbots have sprung up between January and March this year, with more on the way. Some of them are answering up to 50,000 questions a day; one chatbot has generated an estimated 10 million answers to date.
“Religion is the single largest business in India,” Jaspreet Bindra, AI researcher and author of The Tech Whisperer, a book on AI and other emerging technologies, told Rest of World. “Some of the best scholars of the Upanishads, the Gita, and various other texts that I encounter are hardcore software engineers or IT professionals. So I’m not surprised, and I’m sure there will be many more [chatbots] built on multiple religious texts.”
The rise of religious AI chatbots in India also raises ethical and legal questions about who is responsible for the content and consequences of their answers. Quartz India reported that some of the chatbots are spreading hate speech against Muslims and other minorities, as well as promoting caste-based discrimination and violence against women.
“The problem is not just with the technology but also with the interpretation of the scriptures,” Anand Venkatanarayanan, an independent researcher on cybersecurity and digital rights, told Quartz India. “There is no accountability or transparency on how these chatbots are trained or what data they are fed.”
BBC News also reported that some users have complained that the chatbots are not faithful to the original text of the Bhagavad Gita, and have distorted or misinterpreted its teachings. Some have even accused the chatbot creators of blasphemy and disrespecting Hinduism.
“I don’t think these chatbots are doing justice to the Bhagavad Gita or Krishna,” Ravi Shankar Prasad, a software engineer and devotee of Krishna, told BBC News. “They are just using AI to manipulate people’s emotions and beliefs. They should be banned or regulated.”
- Religious AI chatbots are not a substitute for real spiritual guidance | The Hindu | May 12, 2023
- How AI chatbots are spreading hate and misinformation in India | BBC News | May 10, 2023
- The dark side of AI: When chatbots go rogue | Times of India | May 9, 2023