Artificial Intelligence (AI) is a concept that has been around for centuries, and formally studied since the 1950s. In recent years, though, development of AI systems has accelerated, and now major breakthroughs are being made. AI systems are growing in sophistication and scope, and AI is increasingly found in everyday applications – including blockchain and crypto use cases.
AI is the simulation of human-style intelligence processes by machines (and particularly computer software). The aim of AI is to enable computers to sense and process information to solve problems and carry out tasks, replacing humans in domains where these activities might be time-consuming, boring, dangerous, expensive, or otherwise prohibitively complex or undesirable.
Artificial Intelligence covers many different areas, from narrow applications like playing Chess or Go, to broader systems, such as the ability to compose articles (Google’s ChatGPT being a well-known example of this). Ultimately, generalised AI might replicate almost all of the capabilities of the human brain, and be able to learn new skills and engage with a wide range of different activities.
AI is already used in many different applications that are now considered unremarkable – in fact, we don’t think of these as ‘artificial intelligence’ any more, reserving the term for the most noteworthy and new use cases. Just some of the places AI is routinely found include:
Image recognition and phone face unlock software
Recommendation algorithms on Facebook, YouTube, Amazon and other platforms
Chatbots and automated support
Speech recognition technology (including Alexa and others)
Voice assistance and conversational AI, which provide customer service
Financial fraud detection systems
AI art and text generation
Many of these applications are hidden away behind the scenes, and we don’t think much about them any more. The AI processes that provide recommendations and control our social media feeds are just part of our everyday online experience, for example.
However, more complex and generalised solutions are being developed, solving problems and interacting in a way that appears closer and closer to human-level intelligence. Moreover, AI is being applied to new sectors all the time – just one of which is blockchain.
As two cutting-edge technologies, both of which are coming of age, it’s only natural that blockchain and AI should be merged. The rise of AI in Web3 is a notable theme for 2023, with AI blockchain platforms moving into the mainstream, rather than existing as a niche in the industry.
Perhaps the most visible example of this is the articles you’ve no doubt seen in the crypto press, which give a price for a certain crypto on a given date, based on a “machine learning” algorithm. Needless to say, these predictions are anything but reliable.
However, developers are more usefully integrating AI into the blockchain to help users with many tasks such as monetization, data processing, and much more. Tokens that power AI blockchain apps and platforms include:
Fetch.AI (FET). Fetch is a blockchain-based AI and machine learning platform, designed to automate business tasks of all kinds. One example use case is provided by consumer goods giant Bosch, which ran a trial to predict potential failures in their machinery while maintaining data privacy – seeking to create a smart economy of AI-powered devices that automatically meet customer needs. FET is used to pay for network transactions.
The Graph (GRT) is a data aggregation and analysis platform that gathers, organises, and stores information from a wide range of blockchain platforms, enabling other applications to retrieve and use it easily. The software, which is loosely comparable to Google for the blockchain sector, is used by many leading dApps. GRT is used to ensure data integrity and is the platform’s payment currency for the Indexers, Curators, and Delegators who maintain the network.
Numerai (NMR) is an AI-run hedge fund that uses crowd-sourced information and machine learning to invest in global stock markets. Users can submit predictions about stock movements in weekly tournaments, and those who are successful build their reputation – and potential earnings in Numeraire, the platform’s native currency.
SingularityNET (AGIX) is designed to enable anyone to create, share and monetise blockchain-based AI services, with the aim of facilitating the integration of AI into dApps.
AI is capable of reading, parsing, organising, and correlating data at incredible speed, meaning it has a staggering range of applications in the blockchain space.
One of the most obvious examples is analysis of cryptocurrency market sentiment, taking into account on-chain and off-chain price data, social media signals, news articles, and more.
However, the use cases go far beyond trading. Another use case that has already been implemented is supply chain management, including using AI to improve quality, reduce costs, and increase profits for small coffee and cocoa producers.
The combination of AI and blockchain allows applications to process data that is stored on an immutable ledger, and is therefore trustworthy and transparent – ensuring reliability and auditability. As more and more data is stored on-chain, machine learning algorithms will be able to make connections that were previously impossible, providing new insights and efficiencies between sectors.
Artificial intelligence, and AI applications, are now widely used in the blockchain space. As AI becomes integrated into more Web2 platforms, it will logically become integrated into Web3 dApps too.
For example, The Graph supported 20 billion queries in its first year alone, and the technology is now a mainstay of the DeFi space with dApps including Uniswap, Synthetix, Aragon, Aave, Gnosis, and hundreds of others.
Another example of the way that AI is used in the crypto world is in code audits. Certik uses AI-powered tools to secure dApps and identify risks and vulnerabilities for Web3 platforms.
It’s important to state that AI doesn’t run on the blockchain, like the smart contracts that power dApps on Ethereum and other networks. Doing so would be prohibitively slow and expensive in terms of computational resources. But AI software is increasingly using on-chain data as its inputs, and feeding back results to dApps – with an audit trail hosted on the blockchain for security and transparency.
The strength of AI and machine learning lies in the ability to process very large amounts of data and draw useful conclusions from it, potentially making connections between apparently unrelated events. The integrity of the data set is paramount, which is one of the reasons AI and blockchain make such a powerful combination. As more and more data is moved on-chain, we can expect the range of applications that use AI to increase exponentially.
It’s not just the quality of the data that matters, though, but the composability that blockchains offer. The incredible growth of the DeFi sector has been driven by the ability to plug dApps together like Lego, building new applications from existing components. Blockchain AI offers the same functionality, as different data sets and applications can be plugged together to solve new problems.
Once the necessary data is available, many different and disparate data sets might be leveraged to gain useful insights. For example, AI is already being used to attempt to predict certain forms of cancer from shopping habits. With more data about purchases, and more healthcare information stored on the blockchain, machine learning algorithms are likely to be able to draw wider conclusions about a broad range of conditions, and their link with lifestyle factors such as exercise and diet.
Smart cities will use blockchain AI models to predict and avoid congestion, and optimise transport networks for the conditions on a given day, taking into account weather data, traffic volumes, information about accidents, and so on. Businesses will be able to use a similar approach to make sure their supply chains function smoothly, and that they know in advance whether deliveries of materials or components are likely to be disrupted.
Cybersecurity is another area that is seeing significant interest. One of crypto’s biggest problems is fraud. AI is routinely used in fraud-prevention solutions in TradFi, with unusual transactions being blocked until further verification is provided. This is likely to be a major use case, potentially preventing transactions that have cost crypto users billions of dollars in total to date. And, as AI is used to analyse blockchain transaction patterns, it may be possible to track the movement of stolen coins more easily, linking their theft to organised crime and state-sponsored terrorism.
These are just a handful of the many AI/blockchain solutions we are likely to see coming online in the near term. As blockchain and AI become a part of our technological landscape, both will be used to power applications of all kinds – in many cases, without consumers realising they are there at all.