Verifiable oracle protocol ORA launches its on-chain AI oracle (OAO) on the Ethereum mainnet.
While the initial launch of OAO will be on Ethereum, the Oracle will also be live on Optimism, Base, Polygon, and Manta in the coming weeks.
Implementing machine learning (ML) or AI on the blockchain gives machine learning computation access to the verifiability, validity, fairness, and transparency of the blockchain. Despite the benefits, there have been problems with bringing AI into the chain.
First, to enable decentralization, multiple nodes must perform complex machine learning calculations. However, this is quite expensive and time-consuming. Furthermore, Ethereum’s computing environment is specifically designed for EVM smart contracts and is not necessarily compatible with machine learning and AI-related computing adaptations.
Read more: What can blockchain do for AI? Not what you heard.
ORA’s solution to this problem is through its OAOs, powered by optimistic machine learning (opML) on Ethereum. OpML can design any ML model on-chain, although the caveat lies in weaker security assumptions.
According to ORA’s documentation, opML uses a “verification game” similar to that of optimistic rollup systems to enable a decentralized and verifiable consensus on the machine learning service.
Once a requester initiates an ML service job and the server completes it, the results are recorded on-chain. A verifier must then validate the results, similar to what is done in an optimistic merge. If the results are inaccurate, a dispute game starts with the server and the claim is sent to a smart arbitration contract for resolution.
Read more: Funding round: Decentralized AI is all the rage
Grok, a generative AI chatbot; Stable Diffusion, a deep learning text-to-image AI model; and Llama2, Meta’s open source large language model, are now available through ORA’s OAO.
Kartin Wong, the founder of ORA, noted in a press release reviewed by Blockworks that “only opML can put Grok on-chain. This is the supremacy of ORA overML.”
Another type of ML that has been experimented with in the chain is zero-knowledge ML (zkML). This type of technology hopes to generate cryptographic proof for ML calculations that can be concise enough to be verified on-chain. However, current computing power cannot practically generate evidence efficiently and affordably.
Wong claims that opML, unlike zkML, can efficiently bring Grok’s 314 billion parameter model onto the chain, reducing overhead costs by more than 1,000,000x.
