Centralised AI could create a ‘god’, warns Hashed CEO

The future of artificial intelligence hinges on a radical shift: opening the “black box” of centralised models and building a decentralised, transparent ecosystem on blockchain. That is the view of Hashed crypto venture fund CEO Simon Kim, who spoke to CoinDesk.

He argues that unchecked centralisation of AI risks creating a “god” we do not understand. Distributed-ledger technology, he said, offers tools to reward creators, protect intellectual property and ensure transparency for generative models.

“AI is becoming centralised. OpenAI is closed, it is controlled by a small group of people, so it is quite dangerous. Creating this kind of [closed-source] foundation model is like creating a ‘god’ whose workings are unknown,” he said. 

Open-source solutions such as Meta’s Llama show the potential of AI built on decentralisation and transparency, Kim said. However, there is a problem: the lack of robust incentive mechanisms for data providers.

“Models just scan original content on the internet and give an answer without compensating the authors,” the Hashed CEO added.

To address this, Kim proposed an “authorship layer” where rights holders can track AI’s use of their content and receive rewards. He cited Story as an example of such projects.

In November, former Binance head Changpeng Zhao (CZ) called for building artificial intelligence products that use blockchain technology. In his view, the two areas are well suited to work together.

In December, venture capitalists expressed skepticism about integrating AI and cryptocurrencies.

That same month, the DeFi project Frax Finance began developing a technology stack for blockchain-based AI agents in partnership with IQ.