Blockchain and Artificial Intelligence (AI) are arguably two of the biggest drivers of technical innovation in our day. As with previous technological advancements, these two fields are quickly building upon their past achievements (and in an accelerating fashion). Both herald an unprecedented rate of technological development far beyond their fields – and in the very near future as well.
What is blockchain?
If you’re reading this blog, you likely understand what blockchain technology entails. For those that don’t, a brief explanation will suffice. In short, a blockchain is a decentralized public ledger that chronologically records transactions and related information. Each transaction block is irrevocably validated in consecutive order, thus ensuring that the information recorded is unalterable and resistant to tampering. This immutable process is made possible through a variety of network consensus methods, from Proof of Work to our very own Proof of Signature. While cryptocurrency enthusiasts typically view blockchain technology strictly in terms of currency transfer, it also provides the basis for disintermediating contractual enforcement via “smart” contracts. With top-tier engineers, scientists, and businesses leaders flocking to advance both, blockchain technology promises to disrupt a wide variety of industries soon.
What is AI?
Few technologies have garnered as much attention (or captured our imagination as much) as AI. AI, or Artificial Intelligence, denotes computer systems able to perform tasks that typically require human intelligence. Achieving this capability has been an ongoing desire since the advent of computing, from the extremely specific-purpose chess-bot to Google’s new AI Personal Assistant. Nonetheless, these narrow purpose AI tools will eventually give way to True AI or Artificial General Intelligence (AGI), machines that can successfully perform any human intellectual task.
True AI could well be the most important invention in human history, significantly reducing the time needed to perform everyday tasks and conceivably creating a post-scarcity economy. While this scenario is still a while off, AI is moving in that direction. However, its evolution will depend upon accessing a trusted source of data by all. Which is why blockchain technology may dramatically increase the benefits AI brings to the general public.
A Decentralized AI
By using blockchain technology, scientists can record artificial intelligence memories within an immutable ledger. Doing so will ensure that the field retains reliable operating parameters. It will also increase security for AI, ensuring more than one instance of memory, with maybe a backup location or two located in a traditional database. This singular feature
However, even the best artificial intelligence programming is hostage to the quality of its data. As the saying goes, garbage input leads to garbage output. While the blockchain doesn’t ensure access to high-quality data, it does ensure that such data is unalterable. Recording AI programming on the blockchain will not only help to eliminate single points of code failure, it will also provide AI professionals with a source of accumulated knowledge. Sure, the movies tend to sensationalize and oversimplify AI, most notably as a data tool easily altered to wipe out humanity. However, blockchain-secured programming would make it fundamentally impossible to modify AI.
Of course, developers would carefully need to vet such programming before placing it on the blockchain, as it would be immutable and used for encoding additional AI memories. And while validating future AI data (and revising it when necessary) might be problematic, we can certainly ensure that such data is safe from nefarious use.
mprovement and Machine Learning
Recording AI on the blockchain promises to help us better understand why the technology acts as it does. All that is required is for AI to record its decision-making processes as accessible memories. If everyone is using the same base data, researchers will be better equipped to deconstruct AI decision-making, thus helping them predict how AI might handle new situations. This doesn’t just apply to the AGI discussed above, as simple machine learning tools could be significantly improved by analyzing their decision making as well. AlphaGo Zero, for example, is programmed with nothing but if-then rules and an ability to advance against itself. However, it has shortly surpassed any recognizable level of skill through simple self-learning. Now imagine that capability coupled with significantly increased computing power generated from around the world, openly available for other AI programs to learn from and improve. As you might imagine, such a development would be extremely beneficial to the AI field. Of course, as the real world does not typically follow clearly defined rules, advancing AI development in this manner might prove harder than imagined. Nonetheless, researchers believe that continuous machine learning will eventually lead to our first AGI. Placing such machine-learning results on the blockchain could prove invaluable for future AI development.
More importantly, a blockchain based solution for disseminating AI also allows for distributed ownership and control, enabling anyone able to help AI run and to vote on possible changes. This capacity may serve as a desirable safeguard in a future where AI is more or less running the world. A not insignificant achievement given that prominent individuals in related fields believe AI to be one of the more significant threats to our society. Indeed, AI has the potential be a terrifying weapon in the wrong hands. Being in the right hands, it may well thrust humanity into a new age of enlightenment. No matter AI’s advantages, the technology will only be implemented if society can trust it. Fortunately, the decentralized blockchain offers a way out.
Of course, this all remains hypothetical, as such AI has yet to be realized (as is a blockchain capable of running AI). And yet, it’s a necessary technology to consider. The ramifications of its use and future development are extraordinary. For now, however, blockchain enthusiasts are well advised to support continued AI experimentation, particularly as it relates to distributed computing and the blockchain.