Hello, Habre! My name is Anatoliy Krestenko, I am a teacher and mentor at the “DS Specialist” course at the Praktikum, and I have also been working in startups at the intersection of data analysis and blockchain for several years.

In this article, using examples, I will talk about the trends in the combination of AI and blockchain – technologies, the joint use of which was considered impossible for a long time due to conceptual differences.

History of the question

Blockchain and artificial intelligence have long been considered incompatible technologies, both in theory and in practice.

Blockchain is transparent, secure, but known for relatively low data processing speed. At the same time, AI can process large amounts of data very quickly, but it is controlled by corporations and partly remains a black box for business.

Due to technical and conceptual issues, the technologies developed in parallel for a long time and hardly overlapped except for a small number of projects: The Graph (GRT), Fetch.ai (FET), Ocean (OCEAN), Alethea ALI (ALI), SingularityNET (AGIX).

Example, Fetch.ai. It is a platform for developing blockchain solutions based on machine learning and artificial intelligence. It combines Internet of Things (IoT) devices with algorithms for their collective learning.

Several dozen projects at the intersection of blockchain and AI have been launched since 2017, but investor interest in them has been low, and similar initiatives have faced criticism from other market participants.

Community opinion

Opponents of projects at the intersection of AI and blockchain a few years ago were much more numerous than supporters. Participants of the crypto market, whose opinions were listened to, had three main complaints about the combination of technologies:

Prior to the launch of POS EVM networks with a new consensus mechanism, existing blockchains were quite slow. This is the exact opposite of the idea of ​​AI that processes large amounts of data.

Michael Casey, chairman of CoinDesk’s advisory board, also spoke about the problem of scalability: “One of the main problems of combining blockchain and artificial intelligence is the difficulty of scaling. Blockchain has to support a decentralized network — it becomes very difficult and expensive to perform large-scale operations like the processing of large volumes of data required by AI.”

AI is a black box, and the blockchain should be transparent, without any uncertainties. In the blockchain, everything must be clearly calculated and cryptographically confirmed.

This is how cyber security consultant Sarah Lee spoke about this problem: “Artificial intelligence uses large amounts of personal data to learn and make decisions. Often these data are used without the knowledge of their owners. Blockchain, on the other hand, strives for transparency and publicity of data. This paradox creates difficulties in ensuring confidentiality and data security in systems built on the basis of both technologies.”

Machine learning algorithms require a lot of resources to train and operate. Only large corporations can afford it, which are unlikely to agree to conditions in which servers work distributed, and the business receives nothing for it. Blockchain, on the other hand, is based on the idea of ​​decentralization and does not have a single network management center.

Changing the narrative

Although some conceptual issues remain relevant, major changes have occurred in both fields over the past year.

Blockchain projects that experienced a growth boom in 2020-2022 are now experiencing stagnation. The crypto market went through a series of major upheavals: the closure of the LUNA project, the fall of the largest FTX exchange, and the bankruptcy of American banks in which crypto projects kept their assets. By the end of 2022, the crypto market had contracted, and its participants were looking for new technologies to help it grow in the same way that DeFi and NFTs once did.

The same period saw a boom in LLM models, with OpenAI launching ChatGPT, and crypto investors, big players, and developers paying close attention to AI. The index below shows the average relative token value growth of the Fetch-ai, The Graph, SingularityNET, Ocean Protocol, and Oasis Network projects starting with the launch of the model.

By November 2022, several blockchains have emerged that are many times faster than Bitcoin and Ethereum, and some even compete with traditional payment systems.

For example, the Solana network can handle up to 50,000 transactions per second, while VISA has a maximum throughput of 25,000 transactions per second. Ethereum before Ethereum 2.0 could only handle 15-20 transactions per second. Although the market response cannot be called symmetric, the speed of the blockchain to interact with AI has been an important stop factor for the development of the market.

Big projects

At the same time, big players joined the idea of ​​combining blockchain and AI: Mastercard announced the launch of Crypto Credential – a set of common standards and infrastructure to confirm the trustworthiness of interactions between consumers and blockchain projects.

The change in the narrative is evidenced by research, released by Chainlink, the largest blockchain data company. In particular, it talks about the areas of application of AI and blockchain: authentication, security, auditing of contracts, improvement of user experience and interfaces, data analysis.

As an example, Chainlink cites the field of cyber security. In LLM projects, the blockchain can protect the algorithm from unauthorized access: unlike centralized systems, where a single point of failure can compromise the entire system, a decentralized infrastructure is distributed across multiple nodes and multiple independent private keys, making it difficult to compromise the entire system.

This point of view support and at IBM: in addition to security, the company sees the future of projects at the intersection of blockchain and AI in medicine and the organization of supply chains.

Conclusion

In July 2023, Open AI founder Sam Altman launched World Coin, a World ID “digital passport” project. It allows you to prove that the owner is a real person and not a bot with artificial intelligence.

To receive a World ID, a customer must scan their iris using a Worldcoin sphere, a silver orb about the size of a bowling ball. As soon as the scan confirms that the person is real, a record of their World ID appears in the blockchain. At the time of publication of the article, the market capitalization of the project, according to data CoinMarketCap was $270 million.

The number of projects at the intersection of blockchain and AI, according to forecast Spherical Insights will grow rapidly in the coming years. Analysts forecast the Blockchain AI market cap to grow from $230.10 million in 2021 to $980.70 million in 2030. Given the interest from the big players, the field can grow even more actively.