AI and blockchain technology are on the cutting edge of innovation at the moment. Both technologies are making big waves and are showing no signs of receding. While AI has been a field of research for a couple of decades, blockchain recently became popular due to its use in Bitcoin. The popular ChatGPT and Bitcoin, which you have probably heard of, are poster child for how these technologies are changing the world.
AI’s ability to carry out activities that would typically need human cognition has opened up lots of opportunities and is used in a growing number of applications that include online shopping, web search, personal digital assistants, language translation and autonomous cars. This uncanny ability to accomplish tasks that usually need human intelligence has been a source of excitement and fear to many people who wonder how this innovation could change their lives.
Blockchain’s design for security, immutability and decentralization makes it a great replacement for our centralized systems of trust. Some of its profound applications in financial services, web3, art, healthcare, and digital currencies are mainstream today. And a lot of experimentation and tinkering is going on to realise its full potential.
Though, these two technologies have mostly been utilised independently. Ideas are emerging around how they could be used together to create solutions that benefit from their joint qualities. The possibilities that this synergy could bring about are limitless and has the potential to change the course of humanity.
In line with this, this article talks about the opportunities and challenges that exist in the integration of AI and blockchain. It also talks about some real-world projects using them together.
What is Artificial Intelligence?
Artificial intelligence involves the use of large datasets to train computer algorithms to develop models that can effectively carry out tasks that would normally need human intelligence. AI models make decisions based on the patterns they infer from their training data. Oftentimes, the outcome of their decisions is on par with human intelligence. Other times, they exceed it.
AI can be reduced to three components which are – data, learning algorithms, and computing power.
Data
AI needs data to train on. Data often contains inherent patterns that are not easily detected by humans. This data is fed to certain algorithms running on computers that can spot the patterns and learn from them.
The kind of data you can supply to an AI algorithm would depend on what you want it to do. For instance, if your goal is to train a model to spot human faces in pictures, the training data would consist of lots of pictures with human faces in them. Also, depending on the type of your learning algorithm, you may need to use labelled or unlabelled data.
Learning algorithms
There are lots of AI algorithms out there in use today. These algorithms are computer codes written in programming languages and are fed relevant datasets from which they develop and train AI models. Different algorithms are useful in solving different problems. AI algorithms typically fall into one of these categories; supervised learning, unsupervised learning, semi-supervised learning and reinforcement learning.
Supervised and unsupervised learning involve the use of labelled and unlabelled data respectively to train AI models. Semi-supervised learning makes use of initially a small number of labelled data and then a large number of unlabelled data to train its AI model. In reinforcement learning, an agent is set in an interactive environment where it learns by trial and error using the outcomes of its actions as feedback to improve its model.
Computing power
Computing power is needed to run these algorithms that crunch data. These algorithms may take minutes or even months to run depending on the complexity of the problem and how large the dataset is. In some scenarios, very expensive and extremely fast computers called supercomputers are needed to run these algorithms. This is often the case for academic research labs and big corporations running large projects.
The amount of computing power necessary to run an AI algorithm depends on factors such as the complexity of the problem as it is presented in computer code, how large the dataset is and the computing hardware used amongst other factors. Computation is an energy-intensive activity.
What is Blockchain?
A blockchain is the interconnection of many computers to share an identical copy of a database that contains different information on activities carried out by users interacting with it. Blockchain is designed with security in mind.
All the computers on the blockchain network have an equal level of authority, hence, authority is decentralized. The computers work together to agree on what information to add to the database they hold and what version of the database is authentic. Also, the network is open for new computers to join and partake in their agreement activity.
A typical blockchain is made up of a data layer, network layer, consensus layer, application and contract layer.
Data layer
The data layer consists of the data that makes up the blockchain. A blockchain is made up of blocks linked together in historical order. A block is made up of a block header that stores meta-information about the block and a block body which holds several transactions. Transactions contain information on activities carried out by users on the network.
Network layer
The network layer is made up of a peer-to-peer connection of computers (called nodes). These computers jointly maintain the blockchain and are responsible for verifying transactions and adding blocks to the blockchain.
Consensus layer
The consensus layer consists of the consensus algorithms used in coordinating and governing the network. The computers in the network follow the consensus rules to validate transactions and add blocks to the network. Proof-of-work and proof-of-stake are two popular consensus algorithms in use today.
Application and Contract layer
The application and contract layers are made up of rules and instructions written in the platform-specific programming language and stored on the blockchain to be executed when certain conditions are fulfilled. It also includes the user interface through which users interact with them (known as a decentralized application). APIs and frameworks interfacing with these features are part of this layer too.
Potential benefits of their integration
Having described what AI and blockchain technology are, some benefits of integrating the two technologies are
Decentralized AI
Using blockchains in artificial intelligence could help break up data silos and encourage the decentralized storing and sharing of data needed to train AI models. This would allow network participants to take part in transparent interactions and cooperation through the platform. Also, the authenticity and provenance of data are easily verifiable which could help in understanding the decision-making process of AI models. Certain data such as medical data and data from IoT devices which need to be private can easily be kept safe from public scrutiny on this network. Also, the nodes on the network could pool together their computing power to train AI models.
AI-powered Blockchain
The data stored on a blockchain can be analysed using AI to figure out potential faults, possible failures, performance issues and even malicious behaviours on the blockchain. The insights obtained from applying AI could be used to improve the operation efficiency of the blockchain. AI could be used to detect bugs and vulnerabilities in smart contracts and to automate their creation.
Supply Chain Management
Formerly paper-based data could be stored on the blockchain and then fed to AI models to help improve supply chain operations. Blockchain would make it easy to track and trace raw materials and goods for their provenance, logistics, history and other useful information by stakeholders are part of the supply chain.
Health care
Blockchain could be used to improve the interoperability of and access to electronic health records across health institutions. Health information on the blockchain could be analysed with AI to produce insight into public health conditions, personalised health recommendations and so on.
Accounting and auditing
Blockchain could help preserve the integrity of financial data by providing traceable auditing that can be fully automated. AI could help auditors review data stored on the network efficiently by detecting anomalies and evaluating risks.
Smart energy
Blockchain could be used to connect various renewable energy prosumers (a portmanteau of producer and consumer) by creating a decentralized energy market enhanced with automation, security and privacy. AI could be used to optimise the energy network operations to achieve certain goals such as saving electricity bills, improving profits, reducing carbon emissions, effective allocation of energy and so on
Agriculture
The combination of AI and blockchain could be used in predicting crop yields, soil parameters, irrigation requirements, forecasting the weather, identifying diseases and so on. While blockchain could be used to hold agricultural data gotten from IoT devices on farms which AI could analyse to provide useful insights for farmers.
Real-world projects using them together
Here are some live projects that use AI and blockchain today
Bext360
Bext360 provides AI and blockchain solutions in supply chain management for farmers and ranchers so that consumers and other stakeholders involved in the supply chain can trace the provenance, logistics and history of farm products.
Blackbird ai
Blackbird ai uses blockchain to store indisputable and authentic media content from verified sources. AI is used to weigh the veracity of news content and spot misinformation propaganda.
BurstIQ
BurstIQ has a healthcare solution called Lifegraph that securely stores patient data on a blockchain and derives personalized healthcare insights from this data using AI. Patients have access to and control over their data.
Hannah Systems
Hannah Systems use AI and blockchain to provide autonomous vehicles with actionable insights based on their data for a better road driving experience.
Verisart
Verisart helps creators certify NFTs of their work using AI and blockchain so that they get the rightful credit and benefit for their work.
SingularityNet
SingularityNet provides a blockchain-based marketplace for AI tools and services. Developers create and offer AI tools to other users through the platform. The platform coordinates access to these tools using smart contracts.
DeepBrain Chain
DeepBrain Chain combines AI and blockchain to provide a decentralized AI computing platform for AI development.
Some challenges and limitations
Much of the challenges facing the integration of AI and blockchain can be attributed to the current state of blockchains. Blockchains are a relatively new technology and as such are going through a lot of development to reach their peak potential.
Blockchain scalability needs to be dealt with because today’s blockchain transaction processing throughput is relatively constant which implies that it doesn’t scale with an increasing number of transactions. Besides this, delay in transmitting newly mined blocks to the network also affects scalability.
Protection of sensitive data and user privacy on a blockchain are challenges that need to be worked on. Given that blockchain is transparent and open by design, anybody can access this data.
Conclusion
Earlier in this article, we stated that artificial intelligence and blockchain technology are at the cutting edge of innovation and have mostly been used independently. It is evident by now, that both technologies can be used together to create innovative solutions. However, challenges like blockchain scalability need to be fixed for seamless integration to occur.
If you would love to know more about this topic, there are several videos and articles on the internet that talk about the novel applications and possibilities of combining these technologies