Artificial intelligence and blockchain are two of the most discussed technology trends in the world. There are good reasons why these technologies get so much attention, but they are not the only ones.
AI promises to automate many tasks and is often better at modelling complex situations than humans. The combination of blockchain and AI may sound like a counterintuitive pitch for the next big scam. Blockchain provides more data security and privacy, while reducing the overhead and centralised power of large institutions.
But there is a real reason why these two technologies could work well together: the ability to reverse bad trends and improve the quality of outcomes.
As you will see in this article, the AI of the future may well be based on blockchain data sets. Blockchain, in turn, could use AI to monetise user-driven data, create marketplaces for AI models, and even create autonomous organisations. Distributed computing is based on blockchain innovation, but it is also a key component of AI.
In this article, we will examine how these two technologies are currently interacting, and in the decades to come, there will be many developments that will revolutionise the way we think about and talk to these technologies. But it should not be forgotten that AI and blockchain are still in their infancy.
The combination of the two will not be without its challenges, but the ability to deploy blockchain in a wide range of applications, such as smart contracts and decentralised organisations, makes it a viable solution for many applications. Moreover, it is much easier to conclude a smart contract for a decentralised organisation than to write an AI that can safely do things on its own.
Blockchain and artificial intelligence are oil and water in many ways, but we can forget about that, and the potential opportunities can be exciting once we have reached them. In addition to discussing possible applications of blockchain vs. AI, we will also discuss why it is so difficult to mix the two.
Both work in such different ways, but their philosophical opposites are also a major hurdle in the development of new applications. The contrast in philosophy makes the potential of this combination so alluring, and for many of these purposes there is a philosophical contrast.
What option should I choose?
As a result, the companies with the most data and resources will benefit most from AI, which will in turn enrich progress in AI and provide more resources for investment in building better AI. AI relies on training the algorithm to create a place for power in the system, and that place is in the blockchain.
As such, AI is a centralising technology, and consolidating data and computing power encourages the development of more powerful and efficient AI systems, such as machine learning and artificial intelligence.
Blockchain could address some of the philosophical challenges of artificial intelligence, and a common concern is that artificial intelligence makes the rich richer and inherently creates income inequality for the poor that they cannot use.
Blockchain ledgers can decentralise control over data and computing resources, while still making them available to the entire network. Of course, decentralisation comes at the expense of network latency, and serious efforts must be made to accelerate blockchain ledgers if they are to be used for AI.
However, the fact that consumers still own their data and computing power and can rent it out too large corporations if needed is promising. If user data is stored on a blockchain, it can be used by small businesses, governments, or even NGOs.
The result is that a blockchain could democratise access to AI by allowing anyone to develop and use AI models with a real global user base. Access to decentralised computing power would also allow for the creation of new models and the ability for anyone in the world to operate an AI model.
Should we make the data transparent?
The philosophical difference between AI and blockchain is how they deal with transparency. The founding principle of blockchain was transparency, and it is also one of the most important aspects of AI.
An open peer-to-peer network is generally open to all parties that use publicly understandable cryptography. A public blockchain list is available to anyone, and the data is anonymised, but it is still available to those who register. Blockchain is transparent about how transactions are added, so the list of transactions is publicly available.
By contrast, machine learning algorithms and neural networks are notoriously difficult to understand. Blockchain, by contrast, is open, transparent, and accessible to all parties, regardless of age, gender, race, ethnicity, or religion.
While certain types of statistical algorithms require advanced understanding of linear algebra, other deep learning techniques are black boxes even for researchers who have developed them. The transparency problem of artificial intelligence is not a problem that blockchain can solve.
Future research may lead to a better understanding of these algorithms, but for the moment it is simply a matter of further research and development of the technology itself.
Public and independent audits of training data are key to fair AI, and storing blockchain data could facilitate this. After all, an AI model is only as good as the data it was trained on. Blockchain can solve the problem of trained data for AI models that can provide a more accurate and reliable representation of the real data set.
What does the blockchain and AI marriage look like?
Blockchain and AI are philosophically moving in opposite directions, with AI focusing on fast and complex insights based on massive data and computing resources. Blockchain can provide data, computers, and resources, but it will be slower, more transparent, and more decentralised. Speed is the bottleneck right now, and blockchain – scaling solutions currently being developed could deliver data faster.
In combination, these contradictory but powerful technologies could offer several interesting applications. In general, applications focus on democratising the benefits and access to AI, in particular the use of blockchain as a platform for AI research and development.
Feeding AI’s lust for computer power
Algorithm training in AI (machine learning) requires intensive CPU use to calculate thousands or millions of training sessions that teach the algorithm how to make better decisions. Waiting for algorithm training is one of the most difficult and expensive parts of the AI training process. The idea of mining networks is used by the leading possible combinations to obtain computing power as a way to train algorithms.
If data scientists and AI researchers have access to more computing resources, they could develop algorithms faster and more efficiently. Several start-ups are working to connect AI researchers to the thousands of GPUs that are currently promoting cryptocurrencies. Sometimes they can pay the same as miners earn from mining cryptocurrencies, sometimes even more.
The same GPU networks drive the next generation of algorithms and training, so it might make sense to work with them. If the training networks are successful and affordable, they could provide access to the training algorithms for everyone.
If you want to develop a serious AI algorithm today, you do not need access to a central server farm. With decentralised GPU mine networks, anyone anywhere in the world can access the supercomputers to train their algorithms.
I want my Data to stay Private
Although there is still no way to determine the person to whom the data belongs, anonymisation has made blockchain registers a great place for research. By capturing the data collected by Google and Amazon from millions of users, anyone who has access to a blockchain register can use the anonymised data available there to create analyses, make predictions, and even train algorithms.
By training its algorithms on global, anonymous blockchain data, Google could enable it to develop algorithms that are more representative of the world’s population as a whole. The wide-ranging decentralised data available on blockchain registers could also reduce distortions in the algorithms trained on that data.
As we democratise access to computer data, AI is beginning to become an opportunity for collaboration – a development driven by development. Blockchain – data computing based – could break this trend and create a marketplace for algorithms. Companies could pay data scientists for the results of their algorithms or buy access to algorithms for ongoing use. We have already seen the rise of AI and blockchain as a tool for data mining and analysis.
Instead of internal and proprietary, artificial intelligence development could be an open market in which everyone can participate and add value. This type of marketplace is likely to encourage innovation, as we have seen in the development of open source software and the creation of new business models for businesses.
The Data Marketplace
A side effect of blockchain data sets is that users gain more control over their data. Google currently stores more than 1.5 billion data points on its servers and uses this data to create more value and make more money. This data is shared free of charge, but Google currently stores only a fraction of the data available on the Internet, such as Google Docs and Google Drive.
This opens the door to selling access to personal information, but access could be revoked at any time. Maybe you could share information with Google in exchange for personalised services, or a blockchain-based infrastructure could give you control over your data. Ultimately, however, Google would not own the data you share, and this opens the door for you to sell your access and personal information.
This marketplace could provide online users with a passive source of income based solely on existing, saleable data. Businesses could pay consumers for information or gain insights from consumers, and companies could benefit from selling information.
Welcome the Birth of the decentralised autonomous organisations (DAO)
The idea of an autonomous organisation is one of the most important trends in today’s technology world, and not just for the next decade.
There are already many examples of decentralised autonomous organisations (DAOs) in the world, such as Google, Facebook, Twitter and many others. These organisations do not make independent decisions, but only follow the rules of intelligent contract developers. The rules are firmly encrypted in smart contracts so the organisation can take action.