Once Peter Thiel, the co-founder of Paypal, has drawn analogy with crypto and AI from political regimes, naming blockchain-based technology as libertarian and robotics, machine learning and artificial neural networks as communist ones.
It is a controversial question whether these opposites can coexist together or not, nevertheless, by 2018 there are some startups that try to implement both these features and meet crazy public and investing interest.
For example, in December 2017 the SingularityNet, future decentralized marketplace for different types of AI, has raised $36 million in a minute, becoming the fastest ICO in the world. Singapore company Electrify has raised $30 million on ICO, promising to provide electricity on blockchain with AI implementation by the end of 2018. In April 2018 the DeepBrain Chain Foundation has announced its plans to establish AI and blockchain Research Center in Silicon Valley with $100 million of investment into this sphere during three following years.
The possibilities of this high technological collaboration can already be defined and they are likely to change actual perception of business reality.
Contradictions between researchers
In 2017 the predictions about blockchain and AI future were rather optimistic. According to PwC, by 2030 artificial intelligence technologies could add over $15 trillion to the world economy. The blockchain implementation could also bring $3,1 trillion (in comparison with cloud market capitalization that is promised to pass $500 billion by 2020!) to industrial development, as it was predicted in A Gartner Trend Insight Report.
In 2018 the forecasts became much more bleak. According to Forrester Research Inc. investigation, in 90% of cases the companies’ experiments with distributed technologies won’t become a component of real companies’ proposals. Notwithstanding, those who won’t give up on technologies implementation, are likely to be rewarded in future. “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run”, Amara’s Law is provided at the end of the Forrester investigation.
Opposing nature of AI and blockchain
As Reid Hoffman, American entrepreneur, venture capitalist and LinkedIn founder, said during a public debate at Stanford University’s Hoover Institute: “Crypto is anarchy, AI is the rule of law”. According to these words, the general idea of AI is about control and concrete regulations, blockchain is much more about freedom. This explains the centralized nature of the first component and decentralized – of the second one. The unpublic “black box” as a central component of roboethics and AI in general is opposed to transparency and public access to blockchain network.
Black box in AI is a loose concept that includes the approximate understanding of how the machine learning algorithms make decisions. The Sentient company proposes the following understanding of AI black box:
“It is the idea that we can understand what goes in and what comes out, but don’t understand what goes on inside.”
Striking a balance between these categories is not an easy deal.
Dual perception of AI and blockchain
Notwithstanding, both these technologies are frequently met with a good pinch of salt. The current level of AI industrial development in business-to-customer sector is mostly connected with self-driving vehicles, which are followed by “death” scandals with an autonomous Uber crash and Autopilot Tesla accident. The implementation of AI into business-to-business sector comes from conveyor production on the Ford factory and right now is widened by machine learning algorithms, neural networks and experiments in the robotics sphere, for example, with Sophia robot and a great agony around it.
Blockchain, for its part, is a relatively new instrument on the market that meets some challenges not only due to its perception but also because of its technological imperfection.
The failure to achieve real decentralization now, complexity of implementation, current unscalability, that doesn’t allow blockchain to accumulate all the data and make transactions faster (now it might take than a day to be sure your transaction is completed and achieved the goal).
For example, in the context of parallel adoption of both AI and blockchain the problem of scalability can seriously complicate the process of data analysis made by artificial mechanisms.
Mutual benefits of AI and blockchain
- According to an article from Forbes, on one side blockchain technology implemented into AI startup can be a perfect auditor that records all the decisions. On the other side, AI can be managing blockchain functioning even better than human beings do. Machine-learning powered AI mechanisms are an alternative variant to hashing algorithms and their brute strength made by humans. For example, the mining process managed by hashing algorithms is about trying all the possible character combinations until the one that verifies the transaction is found. AI technologies in this case will independently search for the most optimal options, using their black box and solving the problem with less efforts that people do.
- Both AI and blockchain technologies are about functioning of database. If AI needs data to make decisions based on statistics and probability, immutable blockchain is a secure way of monitoring this data and supporting its transparent and “unbreakable” state. It means the blockchain implementation is likely to protect information and make the system independent from any outside human impact. With blockchain protection AI won’t be affected by subjective opinions and it will make decisions that are better for majority, not for private interests of elites, stakeholders and powers that be.
- As Hackernoon’s writers put it, blockchain is a way of democratising AI. This statement seems to sound broad and for everyman it is more obvious that the reduction of costs and increase of efficiency with blockchain application is more about future successful development of technologies than about current situation on the market. But, anyway, blockchain transparency and public access to it indeed can make AI mechanism less about “a riddle wrapped inside of an enigma”. If all the actions are seen, more confidence is shown to both technologies by audience and it will be the first step to democratising of AI. For further “democratising steps”, as one of the VentureBeat’s authors wrote, data quality mechanisms should become more sophisticated. For professional expertise about customers, if we speak about B2C sphere, or about any actions like supplies or court cases, depending on the direction of businesses, high quality of data acquisition technology is needed, then to be used as a basis for machine learning and blockchain implementation.
- Decentralization of blockchain will prevent possible monopolization in the AI sphere. Any AI startup is likely to mean concentration and control in one hands, but implementation of blockchain might nip it in the bud if this technology is introduced throughout. No more conspiracy theories, only worldwide transparency and fair deals as glacé cherry on the cake of blockchain and AI anti-corruption collaboration.
The benefits of AI and blockchain collaboration seem to be a way of evolutionary change in different spheres of business. Therefore, every described technology still needs more technical development to be used at its full strength with the other. In other words, opposites attract, right?