Blockchain and artificial intelligence
Truly effective AI is only as good as the data it is using, the more data, the better, but only if that data is trusted and verified. The distributed network of a blockchain is an ideal framework to enable trust and verification of data whilst maintaining its integrity and security.
The rise of Artificial Intelligence is a complex and controversial subject. Research has been in the forefront of the news lately and it makes for interesting reading. There are however some fundamental issues that need to be overcome before AI can truly take centre stage in business adoption, and with the rise of distributed ledger technology, it seems the perfect storm…. blockchain can help solve some of the issues that have been hobbling the development of workable AI systems in the real world.
AI and blockchain are bedfellows
Blockchain’s cryptographically protected data structure alongside its distributed smart contracts ensures that the data being used can be checked for authenticity before it is processed. If poor or false information is entered within AI systems, poor AI decisions will be made, it’s a simple equation really. Blockchain’s trust protocols are ideally suited to this situation, solving data integrity at a stroke.
Alongside this integrity of data, comes the increased ability to share it with other systems, quickly and efficiently. In fact, the construction and nature of these AI mechanisms are not a million miles away from the decentralised architecture of blockchains.
Modern AI requires its material data to be pushed pulled and merged so that it can run and test many different hypotheses, then the differing possible outcomes will be tested again until the final most positive outcome is chosen and suggested. Instead of running these hypotheses through large centralised, proprietorial systems, it makes much more sense to use the distributed nature of blockchain to run these actions through a system of nodes spread across the globe all securely monitored in a permissioned network. That way the data can move securely and freely without bottlenecks or centralised points of vulnerability.
AI and the real world
Most if not all AI systems being developed currently exist in an environment known as the red ocean, where each individual proprietorial system is contending with other similar systems for individual gain. Each using their own siloed data to compete with the other.
In most industries this is the norm, but where AI is concerned this ringfencing of competing data is an anathema. It restricts the scope and breadth of the data available and thus the effectiveness of being able to apply machine learning is greatly reduced.
A much more positive environment to allow AI to flourish is called the blue ocean (derived from “Blue Ocean Strategy” (Harvard Business Review Press, expanded edition, 2015), by W. Chan Kim and Renee Mauborgne.) Here competition is irrelevant, and data is shared willingly in the pursuit of high efficiency and low cost. It just so happens that blockchain is most effective when positioned squarely within this blue ocean environment. This all sounds very idealistic I know, but if AI is to flourish it requires a re-adjustment in how we see, value and use data on a day by day basis.
In a world where data is hoarded by giant organisations such as Google and Facebook, AI can only empower those organisations with the biggest silos. That does not make for an effective environment for the development of AI. Although these companies are financially able to pump huge resources into developing this technology, history tells us that the real innovation will come from a far more open and outward facing innovation strategy.
history tells us that the real innovation will come from a far more open and outward facing innovation strategy.
Blockchain has rapidly been evolving and the systems available now are able to be managed in permissioned or even private networks where there is control and governance of these distributed networks. These permissioned or private networks also offer speed and scalability that previously were just not possible. Blockchain is becoming cheaper faster, more secure and more practical by the day.
Whose data is it anyway?
Does the data that these big corporations hold really belong to them? As we understand the increasing value and usefulness of this data we begin to question the fact that our data is owned by others, sold by others and accessed by organisations that we have no knowledge of. Data breaches are becoming more commonplace and increasingly these centralised systems are appearing ill-suited and archaic when compared with how our use of data is evolving.
Key to the evolution of AI is this attitude to data. Any business that is exploring AI is more than aware of the vulnerability of the raw material data it possesses and is compelled to invest huge sums in protecting this data against attack. More stringent legal requirements also increase the complexity of how you manage this resource… This hobbles business innovation.
With this increased efficiency comes an increased synergy with AI in data structure and movement alongside an increased benefit of security, transparency and trust. Blockchains have always been seen as an improved way of storing value and transmitting it from user to user, so when you look at data as a valuable resource within the world of AI then the synergies seem obvious.
It seems to us at Omnitude that the future of AI and the blockchain are inextricably linked.