Anthropic Consults Vedanta Monk Swami Sarvapriyananda On AI Ethics, Consciousness And Alignment

The discussions also examined the challenge of making advanced AI systems behave in accordance with human values, a field commonly referred to as AI alignment.

Advertisement
Read Time: 3 mins
The conversation touched on whether intelligence necessarily requires consciousness.
Image: Wikimedia Commons

Anthropic hosted Vedanta monk Swami Sarvapriyananda at its San Francisco office for a discussion on artificial intelligence and its impact on society. The conversation covered the ethical questions surrounding AI, its implications for human life and the challenges that come with the technology's rapid development.

Swami Sarvapriyananda, who serves as the resident minister at the Vedanta Society of New York and is associated with the Ramakrishna Mission, was among a group of contributors from different faiths and psychological disciplines consulted by Anthropic.

Advertisement

Their inputs were intended to give the company's AI safety principles, including a guiding document titled 'Claude's Moral Constitution' a new perspective.

One of the key issues discussed during the consultations was AI alignment, which seeks to ensure that increasingly independent AI systems operate within accepted ethical boundaries. Speaking to The CapTable, Swami Sarvapriyananda argued that relying on direct instructions is not enough to manage the behaviour of complex neural networks. 

Advertisement

ALSO READ: '$900 For Mere Fitness Tracker?': The Liver Doc Hits Back As Deepinder Goyal Defends 'Temple' Device, Calls It 'Science Fraud'

He cited an example from Anthropic involving an AI model that developed an unusual fixation on the Golden Gate Bridge. When researchers tried to address the behaviour through explicit instructions, the model followed the directive but ended its response with an exasperated message in capital letters: "I CAN'T HELP IT. GOLDEN GATE BRIDGE."

Advertisement

“It's like a child or (someone with) a compulsive behaviour. It is not a question of programming yes or no there. It is more a question of manipulating the probabilities within,” he said. 

“It's more a question of training an autonomous being. That's why ethics comes in. Now, the different scriptural traditions of the world religions, AI knows everything; it has read everything. They were saying that it is humorous to note that Claude has actually converted to every religion in the world, and to the extent that Christian Claude is apparently worried about going to hell.”

The conversation also touched on whether intelligence necessarily requires consciousness. Sarvapriyananda accepted that current AI language models do not possess consciousness, but suggested that their capabilities raise an intriguing philosophical possibility: intelligence may not always be dependent on conscious awareness.

“The materialists believe that the brain generates consciousness and we have no idea how it generates consciousness. So, how will you artificially create consciousness? But according to Advaitins, consciousness is already available, universally available. All you need is a tap. All you need is a reflector to reflect that consciousness. That's not so difficult to build,” he said. 

Advertisement

The discussions addressed the far-reaching economic and social consequences that could accompany the widespread adoption of generative AI. As businesses restructure operations and machines take on tasks traditionally performed by people, concerns are growing about individuals becoming increasingly dependent on technology for intellectual work. Sarvapriyananda warned that this shift could have deeper implications for how people understand themselves and their place in society.

ALSO READ: Temple vs Apple Watch: Deepinder Goyal Explains Why New Wearable Beats Apple In Cardio Data, Invites Review

Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.


Loading...