Anyone that’s been following the tech world knows that there many major technologies that are innovating at an exponential rate. The first of which, blockchain technology, is a subject you’re likely to already have heard about. The second, Artificial Technology (AI), has been getting just as much buzz lately in the last couple of years. But what many people don’t know is that the same technology that has ushered in the cryptocurrency revolution has also helped solve a big piece of the puzzle surrounding AI’s future development.
The term “artificial intelligence” was first coined as early as 1955, by Professor John McCarthy. Since then, scientists have now begun to predict that AI technology will outperform humans in many areas, such as translation (predicted for 2024) to even complex surgery (around 2053 supposedly).
But before we begin to explain just how blockchain technology is promising to transform the field of AI, we need to go over some background information pertaining to the field.
The Artificial Intelligence Arms Race
Although commercial applications of AI technology, such as Apple’s virtual assistant Siri, can help people save time on tedious tasks, this same technology on a global level is seen as a key asset for countries that seek geopolitical dominance over their rivals.
The consensus is that once AI has become developed enough, the technology can lead to significant strategic advantages due to AI’s superior intelligence to regular humans. Furthermore, world leaders and AI experts theorize that the first nation to birth a superior AI will have a permanent advantage over its competition. What this means is that the first advanced AI can develop other AI’s, escalating into an exponential growth that eventually no human team will be able to compete in – regardless of whether a nation creates this ‘super’ AI only months or even weeks in advance of its competition.
Labeled as the global AI arms race, billions of dollars have been invested in research and development for these programs. In the words of Russian President Vladimir Putin, “Whoever becomes the leader in this sphere will become the ruler of the world.” Other figures, such as Elon Musk, have warned that competition for AI at a national level is potentially the greatest threat to humanities future existence.
As for how this relates to blockchain technology, think of it this way. If one of the allures of cryptocurrencies is it’s potential to decentralize what is often seen as a burgeoning global financial system, implementing blockchain technology into the development of AI could drastically reduce people’s concerns over the misuse of this technology. Few would like the prospect of this powerful technology being controlled by a single government, corporate, or party – which ties in nicely with our first point.
The Possibility of Decentralized Intelligence
The idea of a decentralized AI distributed over some kind of blockchain helps neutralize the potential risk of one party gaining an unbeatable advantage over the use of AI technologies. If this decentralization process took place, AI algorithms would operate in a fundamentally different way. Decentralized Autonomous Organizations (DAO’s) are organizations that operate autonomously and in a decentralized way through smart contracts without having any single party pull the strings or make decisions.
Participants on a blockchain network are directly connected with each other, and like a variety of network ecosystems out there already, AI programs running as DAO’s would work the same way. In contrast to the popular mainstream idea that an AI can “go-rogue”, an AI DAO operates through the use of smart contracts – able only to perform the actions specified in these contracts.
Of course, designing these smart contracts would be quite complex and the theoretical success of an AI DAO would be dependent largely on how sound said contracts were from an architectural perspective.
Data Distribution and Protection
Any AI’s potential for progress, or even to function properly, is dependant on large quantities of data. A basic law of statistics states that the accuracy of a study or analysis increases as the data set grows larger, and AI technology requires an extremely large data set in order to function accurately.
On the other end, blockchain is a technology that allows for the encrypted storage of data on a ledger distributed across the entire ecosystem. It allows for the establishment of fully secured databases that can be looked into by parties who have received permission to do so. The idea is that blockchain technology can work as a backup system for any personal or sensitive information required by AI algorithms. Any parties that aren’t approved, such as potential hackers, won’t be able to acquire this information. When it comes to financial or medical data, this information can be too sensitive to hand over to a single company and its algorithms. By storing this info on a blockchain that’s accessible by an AI – but only if it has permission and once it’s gone through the securitization procedures – we’ll be able to make sure our information remains secure from outside parties.
Monetizing Our Data
While on the topic of personal information, it’s worth noting that blockchain technology in conjunction with AI developments can change the playing field for how our data is monetized.
It’s well known that data mining is a large source of revenue for large companies such as Google and Facebook. By using the personal details of their users, this data gets collected, grouped, stored, and eventually sold to other organizations. As the 21st century goes on, this data is going to become increasingly valuable – especially once regular users realize that they are essentially giving it away for free on these sites.
Blockchain technology makes it possible for common users to monetize our data, owning our own private information and deciding for ourselves whether we want to sell this info to data marketplaces. As opposed to the current situation, where third parties sell our data to organizations (often using unethical means, as the recent Facebook scandal revealed), companies will have to buy it directly from the end user.
This would also apply to AI algorithms that need our data. Instead of using the large databases of tech and social media companies, AI networks will need to buy data directly from the end users themselves through a data marketplaces, making the entire arrangement fairer for everyone involved.
It would also open up the possibility of AI development to smaller companies. Since feeding an AI continuous streams of data is quite expensive if a company doesn’t have its own methods of data mining, decentralized data marketplaces can give alternative AI developers a chance to secure a constant source of information at a reasonable price.
Verifying AI Decision Making
As AI algorithms become more intelligent through deep learning, it becomes ever more challenging for data scientists to figure out how these AI’s came to their conclusions. Through processing vast quantities of data, it becomes ever harder to audit the decision-making process of an AI and verify that their deductions were correct.
What blockchain technology allows is clear documentation and a list of all the data, variables, and processes an AI used when it makes a decision. All steps from data entry to the final conclusion can be observed thanks to appropriate blockchain programming, and data scientists can also verify that this information can’t be tampered with.
The end result is that people will be able to more fully trust the conclusions drawn by AI algorithms, rather than simply hope that the decision-making process hasn’t gone awry due to some faulty/tampered data.
The contributions of blockchain technology to the world of artificial intelligence promises to shore up some holes in the future application of AI. The combination of these two fields helps address a number of concerns regarding the decentralization of AI as well as mitigate some of the risks concerning future data mining and utilization. While these developments are still a few years down the road, there’s no denying that they will come more quickly then they would have otherwise without the contributions of the blockchain.
Also published on Medium.