By
Temitope Olatunji
April 26th, 2024
The potential impact of blockchain on AI governance goes beyond addressing bias and misinformation. It offers a pathway to ensure the integrity, accountability, and trustworthiness of the data used to train AI models.
During a discussion at the World Economic Forum in Davos, Switzerland, executives touted blockchain, the technology behind Bitcoin (BTC), as a potential game-changer in addressing bias and misinformation in the training data of artificial intelligence (AI) models.
According to a CNBC Monday report, the executives believe that blockchain could revolutionize how AI systems, such as ChatGPT and Google Bard, are developed, ensuring a more accurate and unbiased outcome.
Verify AI Dataset on the Blockchain
Blockchain, introduced in 2009 with the launch of Bitcoin, operates as an immutable, tamper-proof public ledger of transactions. Experts believe this technology can be harnessed to create a secure and transparent system for managing AI training data. The concern with current AI models is the presence of biases and false information in their training data, leading to potentially skewed or inaccurate responses from the AI.
To eradicate this concern, Casper Labs, a prominent blockchain firm focusing on business applications, recently joined forces with IBM to pioneer a system where AI training data is stored on the blockchain.
During the panel discussion in Davos, Medha Parlika, the chief technology officer and co-founder of Casper Labs, elaborated on the project, emphasizing how datasets can be stored on the blockchain, providing a verifiable record of the AI’s training process.
“The datasets are checkpointed and stored on the blockchain, so you have proof of how the AI is trained. If the AI starts to hallucinate – providing false information – you can roll back the AI to a previous version, undoing some of the learning,” Parlika said.
Parlika additionally mentioned that the project will enable reverting AI models to their initial stage in case biases or inaccuracies are identified. “And so as you use the AI, if it’s learning and you find that the AI is starting to hallucinate, you can actually roll back the AI. And so you can undo some of the learning and go back to a previous version of the AI,” noted she.
Killer Use Case for Blockchain
Another executive at the Summit, Sheila Warren, CEO of the Crypto Council for Innovation, spoke of the transformative potential of this integration, referring to it as the “killer use case” for blockchain technology that could redefine the landscape of artificial intelligence for years to come.
In her view, the approach could provide the necessary checks and balances, ushering in a new era where the verification of AI processes is blockchain-driven and blockchain-backed.
“I actually do think that the verification of an AI and sort of the checks and balances … within an AI system, are going to be blockchain driven and blockchain backed,” said Warren.
Different sectors, including finance and healthcare, are currently exploring the use of blockchain in managing AI training data to ensure the integrity of AI applications.
Meanwhile, the potential impact of blockchain on AI governance goes beyond addressing bias and misinformation. It offers a pathway to ensure the integrity, accountability, and trustworthiness of the data used to train AI models.