The Emerging Nexus of AI and Blockchain
Fourth Industrial Revolution technologies have been advancing – and continue to advance – at a staggering pace. What ultimately drives this pace is the new applications and uses being discovered and/or created for these technologies – as rarely does society understand how to extract the full value of a new technology when it is first introduced. Doing so usually takes several years of experimenting with the technology to find optimal uses for it.
This is certainly the case for both AI, or Artificial Intelligence, and blockchain. Both of these technologies are making countless new activities possible in our world that were inconceivable a mere twenty years ago. Even more importantly, the nexus – or marriage – of these two technologies are making possible some incredibly powerful new uses and applications.
The Purpose and Functions of AI & Blockchain
Let us first briefly revisit the purpose and function of each of these two technologies, and then we will peer into the power of their collective nexus.
Artificial Intelligence – of which a variety of types exist, each based on a certain algorithmic methodology – exists to ultimately make useful decisions for us (based on certain heuristics and inputs) and thereafter (at least in the most advanced cases) take certain useful actions for us. It does this by assimilating, analyzing, and interpreting large volumes of data so as to make meaning of that data, and thereby infer a usable insight from the data, from which it then makes a relevant decision, and then – based on that – ultimately undertakes a relevant action – all, ideally, in real time.
Blockchain – a distributed ledger technology – exists to create an immutable record of a chronological series of events – whether those events represent a decision, a change, a transaction, or any other type of documentable ‘event’ of interest. By creating an immutable record of each such event, its evidence is permanent and persistent – with us forever. And on account of the nature of how the blockchains are computed, it is – at present, and given present technological capabilities – accepted as a credible and reliable confirmation of the event. Which is why blockchain provides the perfect – and necessary – basis for such devices as cryptocurrencies and digital tokens.
The Nexus of AI & Blockchain Take 1 – AI Built On Blockchain-Governed Data
Now – in thinking about the nexus of these two technologies – one can understand that there are, and very quickly will be more, situations in which we (society) want automated AI systems to make certain decisions for us and thereafter take certain actions for us, but such decisions and actions are of such a crucial and dire nature that they must be based on clearly verified data. In other words, the consequences of those decisions and actions (some of which are irreversible) are of such a gravity that the data they are based on must be unquestionably verified. Blockchain provides the tool by which to validate the integrity of that data and thereby ensure the decisions being made and actions being taken are all built on reliable, verifiable insights.
Examples of such situations can include any of the following:
- Having AI make business decisions for us – such as those involved in buying materials, selling goods and services, managing supply chains, managing intellectual property, and so on – and thereafter take corresponding business actions on our behalf.
- Having AI make financial decisions for us – such as those involving investing and divesting, or securing credit and loans – and thereafter undertake financial actions on our behalf.
- Having AI make legal decisions for us – such as those involving business partnerships, estates, trusts, wills, and so on – and thereafter take legal actions on our behalf.
- Having AI make medical decisions for us – such as those involving disease diagnosis and treatment, palliative care, and pharmaceutical usage – and thereafter take corresponding medical actions on our behalf.
- Having AI make policy decisions for us – for those setting policies, whether in businesses, governments, or elsewhere – and thereafter undertake policy actions for us.
In all of these situations, the inputs to, and the outcomes of, the decisions being made and the actions being taken must be reliably verifiable – and verified – via blockchains. This ensures the validity of the information being operated on. It can also – with certain additional steps like private blockchains operating on encrypted data – aid in ensuring privacy, such as is a concern in healthcare and patient privacy, or in legal services where client-attorney privacy is crucial.
It also – in cases where Machine Learning is being used – produces better and more reliable AI models because the data on which those models are being trained has been verified. That in turn allows us to better trust such AI models and their resultant decisions / actions. This is crucially important wherever distributed communities are building AI models together via a public blockchain – an increasingly common practice known as ‘decentralized artificial intelligence’.
The Nexus of AI & Blockchain Take 2 – Blockchain-Governed AI Models
One might likewise understand that there are, and will likely be more, situations where we (society) want automated AI systems to make decisions for us and thereafter take actions for us that are of such a crucial and dire nature that they must be based on clearly verified AI models. In other words, the consequences of those decisions and actions (some of which are irreversible) are of such a gravity that the models they operate on must themselves be unquestionably verified. Blockchain provides the means by which to validate the integrity of these models and thereby ensure the decisions being made and actions being taken are all achieved via verified models.
This is the converse of the prior situation, where the data had been governed by blockchain; here the AImodels are themselves governed by blockchain – storing and distributing the models with an audit trail. This enhances even further the security of the decisions being made and the actions being taken – and provides for what is known as ‘explainable AI’, in which the AI decisions are (at least in theory) understandable by humans because each step involved in them has been traced via a blockchain.
Indeed, a powerful and increasingly common use of AI-blockchain is that of blockchain-based smart contracts with embedded AI models. These are used to do such things as:
- execute transactions — like payments, re-orders, and stock purchases based on thresholds; select optimal shipping methods;
- resolve disputes;
- suggest expired products to recall;
- and more.
Companies Developing AI-Blockchain Solutions
There are a number of businesses today developing solutions that effectively merge AI and blockchain together. Examples of some of these include – by sector:
- Industrial: Core Scientific (corescientific.com) / WorkDone (workdone.ai) / Stowk (stowk.com)
- Financial: Figure (figure.com) / WealthBlock (wealthblock.ai) / Numerai (numer.ai)
- Healthcare: BurstIQ (burstiq.com) / Vytalyx (vytalyx.io) / Gainfy (gainfy.com)
- Urban Infrastructure: NetObjex (netobjex.com) / Hannah Systems (hannahsystems.com)
- Entertainment / Content: ObEN (oben.me) / MOBS (mobs.video)
- Energy: Via (solvewithvia.com)
- Art: Verisart (verisart.com)
- Cybersecurity: Cyware Labs (cyware.com)
- Food & Mineral Sourcing: Bext360 (bext360.com)
- Weather Forecasting: Neureal (neureal.net)
- News: Blackbird.AI (blackbird.ai)
All of the above – including these commercially-available platforms – vastly improves the integrity of the decisions being made and the actions being taken, which in turn drives a far more trustworthy and transparent data economy across businesses and governments.
They also make the insights discovered by, and decisions made by, such systems far more valuable, which can have important commercial implications whenever such insights and decisions can be sold commercially.
The Present Shortcomings of AI-Blockchain
The above notwithstanding, there are currently still two major shortcomings of this marriage of AI and blockchain.
The first shortcoming has to do with the time required to process blockchain transactions, which is still measured in minutes, not milliseconds. The implication of this is that AI-blockchain activities, while incredibly powerful, cannot presently operate in their native form in real-time. Thus real-time decision-making and real time action-taking via this combination are still out of reach for us. That could change however, as algorithms and hardware advance further yet.
he second shortcoming only comes into play whenever physical goods are involved, such as when trying to trace food products back to their sources, or pharmaceutical goods through their lifecycle, or luxury goods for customs-enforcement purposes. In each of these cases, the shortcoming lies in the transference of a digital record to an associated physical product. While this can be done via a variety of means, such as with barcode labeling and (more elaborately) RFID-tagging, the fact remains that there is some level of human activity involved in doing this, and therefore some level of traditional human-based trust required. In reality, bad
actors can – with enough effort – intercept shipments of physical goods and switch those goods from the originals to alternatives under the same traceability tags. So bad actors can still corrupt physical systems, despite our good intentions otherwise with blockchain. In the case of physical goods, those records are only as good as the extent to which we believe the goods we are receiving are the same goods the records were originally associated with – with no bad actors intervening in between. This is clearly a weak link in the chain of custody of such goods. This shortcoming does not exist (generally speaking) with digital information, in that it is far more difficult to infiltrate the chain of custody undetected. Thus those using AI-blockchain to track physical products must be aware of this inherent potentiality and its associated risk.
Conclusion – AI-Blockchain Is A Pivotal Combination
Despite the present limitations of AI-blockchain, there remain some incredibly powerful uses for this nexus. These largely revolve around either building AI applications on top of blockchain-governed data, or else deploying blockchain-governed AI models themselves – or ideally both together.
This nexus ultimately results in far more reliable valuable insights, making far more dependable decisions, and consequently taking far more optimal actions for us. This is, indeed, the level of confidence that society needs to be able to scale AI up to massive levels across our business, government, social, political, and other infrastructures around the world.
AI-blockchain is, at the end of the day, the one nexus the world truly needs to be able to realize the full – and massive – potential of AI across every aspect and domain of our lives. It protects us against the malfeasance of bad actors manipulating our AI systems for their own gain, and in so doing gives us the assurance we require to trust AI as a reliable aid in optimizing our world.
Blog Courtesy: Anthony Mills
Anthony Mills is a globally sought–after thought leader in areas such as emerging markets, proactive growth strategies, corporate innovation, open innovation, the future of work, workplace experience, product design, entrepreneurship, and design thinking. His work has had a deeply profound and lasting impact on business organizations all over the world.