Blockchain Could Be the Solution to AI Bias, Executives Say

Artificial intelligence (AI) models are becoming more powerful and ubiquitous, but they also face challenges such as bias and misinformation. To address these issues, some executives suggest that blockchain, the technology behind bitcoin, could be used to verify and track the data that AI models are trained on.

AI models are computer programs that learn from data and perform tasks such as natural language processing, image recognition, and recommendation systems. However, the quality and accuracy of these models depend on the data they are fed. If the data is biased, incomplete, or inaccurate, the AI models may produce wrong or harmful results.

Blockchain is a system that records and stores data in a distributed and immutable way, creating a transparent and verifiable record of transactions. Blockchain was first introduced in 2009 with the launch of bitcoin, a digital currency that operates without a central authority. Since then, blockchain has been applied to various domains, such as finance, health care, and supply chain.

Some executives believe that blockchain could also be used to improve the trustworthiness and accountability of AI models, by creating a ledger of the data that the models are trained on. This would allow the developers and users of the AI models to check the source and quality of the data, and to detect and correct any biases or errors.

One example of such a blockchain-based AI system is the one developed by Casper Labs, a blockchain company that partnered with IBM this month. Casper Labs’ chief technology officer and co-founder, Medha Parlika, told CNBC that their system stores the AI training data on the blockchain, providing a proof of how the AI is trained.

Parlika also said that their system allows the users to roll back the AI if they find that it is giving false or misleading information, which she called “hallucinations”. By doing so, the users can undo some of the learning and go back to a previous version of the AI.

Sheila Warren, the CEO of the Crypto Council for Innovation, a trade association that promotes the adoption of cryptocurrencies and blockchain, said that blockchain could be the “killer use case” for AI. She told CNBC that blockchain could provide the verification and checks and balances that are needed to ensure the reliability and integrity of AI models.

Warren said that she thinks that blockchain will be the backbone of the AI systems of the future, as it will enable more transparency, accountability, and trust in the technology.

Read more:

Join us on Telegram

Follow us on Twitter

Follow us on Facebook

Follow us on Reddit

You might also like