Digital trends of tomorrow — Artificial Intelligence and Blockchain and their business values
The rise of AI in the business world is powered by the combination of growing processing speed, learning algorithms, and the availability of data. AI is currently progressing in the Narrow or Weak segment, and we are far away from reaching working General AI. AI market size is expected to reach ~$50 billion in 2020. Early adopters of AI technologies that can upscale their talent pool and enforce data governance best practices will keep a competitive edge.
Blockchain is a protocol to implement a distributed open ledger that securely keeps record without the need of a trusted third party. Although block chain is at the beginning of it’s journey, it has already applications in financial services, assets traceability, contract management, and many other domains. There are still multiple challenges associated with blockchain from technology difficulties to questions about the business model and the regulatory framework. However, if blockchain makes its way, it can completely disintermediate established platform players in today’s business ecosystem.
Artificial Intelligence
Today, it is undeniable that machines are beginning to build human-like capabilities and in many instances, AI has already proven to be better than humans
The underlying trends: processing power, big data, algorithms
IBM’s Deep Blue that defeated Garry Kasparov back in 1997 used a very simple algorithm. From a given position, the computer evaluated all possible moves and, for each possible move, it evaluated the possible responses from the opponent, and continued doing this evaluation for as many rounds as it could depending on the processing speed and the time available. With such an analysis, the computer simply chooses the move with the best prospects.
The success of Deep Blue was due to the massive increase in processing power between 1996 and 1997. Kasparov won against Deep Blue’s AI in 1996 but lost in 1997 when the machine doubled its processing power. The continuous increase in processing power has led to capabilities beyond chess play. However, processing power alone is not enough to build an AI. At least two other ingredients of Big Data and Algorithms are needed:
- Big Data — we need ways to capture and accumulate data inputs from the world the AI machine is supposed to be interacting with
- Algorithms — we need to find the right algorithms and techniques to process the data input to achieve desirable outcomes
Defining AI
AI is a system that can exhibit traits of human intelligence such as reasoning, learning from experience, or interacting with humans in natural language (e.g. iPhone’s Siri and Samsung’s Bixby). But the definition of intelligence could be rather subjective and changes over time with technological progress. To make this slightly more objective, we can distinguish two types of AI — general and narrow AI.
- General AI — is typically a complete system that is indistinguishable from a human. It knows or can learn anything humans can learn, has emotions, and even has a purpose in its life.
- Narrow AI — is when a system exhibits human-like intelligence traits on a specific field or task. For example when a system knows how to play chess, but can’t write a cake recipe or when it can detect a traffic signal but can’t detect cancer.
Today, we don’t know how to build general AI, but we are pretty good at building narrow AI because its ingredients of sufficient processing power, lots of available data, and the right set of techniques and algorithms are readily available.
The algorithms used for Narrow AI are specific. For example, an algorithm that is capable of looking at thousands of bird pictures and recognizing a specific kind of a bird. This is called Machine Learning. One popular way of writing such machine learning algorithms, inspired by the way our brains work, is called Neural Networks. Neural networks— put in simple terms — take input data — for example, a picture — and feeds it into its artificial neurons that work together to recognize whether there is a specific bird in the picture. Neural networks are a scalable technology whereby we can combine many layers of neurons in what is called Deep Learning algorithms to achieve more complex functions, such as to come up with accurate sentence translation by learning from all the web pages that are available in multiple languages.
Today, we don’t know what will take our technology from narrow AI to general AI, but we have already reached a point where machines can interact with the external world through natural language processing (i.e. Siri or Bixby), where they can decode human faces and understand their emotions.
Business value of AI
Today, most technology players are heavily investing in AI startups, whether through acquisitions, share purchase, or even internal entity developments. In addition to major tech players, global investments in AI have gone from ~$0.6 million in 2012 to ~$5 billion in 2016, and is estimated to reach $50 billion in 2020, a 10X increase within the span of five years. Moreover, +60% of executives believe AI will have a large effect on their businesses in the next five years with industries such as healthcare, automotive, telecom, and financial services leading the way.
This trend of capital movement means that if it is time for all businesses to start experimenting and understanding the potential applications of AI for their specific industries to build competitive advantages.
- Computing power — is becoming more and more a commodity and unless positioned in a very specialized industry, competitive advantage will not be gained here
- Algorithms — although a lively field of research, are also a commodity. The limiting factor with algorithms is finding the right talent, upscaling the workforce, and creating an AI-enabled environment to keep up with the evolution.
- Training data — is the area for the development of competitive advantage and intellectual property.
The challenge with AI
The development of AI will lead us towards serious ethical and sometimes legal decisions and implications. For example, if a self-driving car has a choice between two courses of action, one which will kill one pedestrian, and one where it will kill five, which one should it choose? And what if the one is not a pedestrian, but the car owner in the front seat? Should the car sacrifice the owner for the greater good? And would you personally buy this car? So which path should our algorithms take here?
In most cases, AI algorithms are black boxes that learn from our experiences to make their choices, but no one is able to explicitly identify which input variable or which past experience led to which choice. For example, AI algorithms can develop discrimination biases based on race or on gender, such as when selecting the best candidates for a given job description. Based on previous recruiting campaigns, the AI will likely be replicating the same biases that human recruiters might have against certain minorities, or the algorithm might learn for example that female offers should be lowered than males, because that is still, unfortunately, the case in many companies today.
Some of these dilemmas will likely be solved with improved understanding or better control of AI algorithms, but others will need policy and regulatory changes.
Blockchain
Imagine that you want to buy a house in the countryside. You will probably not trust an online add with some descriptions and photos, but rather first want to make sure that the asset has an owner and is real. Besides, you’d want to have an objective record of what has happened to the house, when it was built, how big it is, its renovation history, etc. Traditionally, this information would be stored in a bookkeeping record or ledger that includes all the asset history. The ledger would typically be stored centrally under a title authority or an intermediary that you’d be able to trust and that facilitates the transaction between you and the owner of the house.
This intermediary authority will guarantee that you are buying the right asset from the right owner, it will keep records of your purchase and you will refer to it later when you need to prove your ownership. It probably takes some time to update and consult the ledger and the title authority will charge you a fee for it. This ledger and authority is potentially the best solution to facilitate the transaction, prior to the entry of Blockchain technology
Blockchain and its core features
Simply speaking, blockchain is a ledger built in a way that allows you to trust its information without needing an intermediary authority. In the house purchase example, it will include all the records on the house you want to buy. But two characteristics make the blockchain ledger fundamentally different from the traditional ones we know today.
- It is open
- It is distributed
Open means that anyone can have access to it and read from it or write new transactions into it, but there is a protocol that accepts only the transactions that make sense. For example, you can only transfer the ownership of a house if you happen to own it. Therefore, to verify a transfer, every record or transaction that is written in the ledger includes a digital signature that uniquely identifies who has written it. The combination of a few transactions and their signatures is called a block. If the transactions within the block are allowed by the ledger protocol, the block is then signed by a unique key that validates the records. When the next transaction is created, a reference to the first block is included at the start of the second black to guarantee that the sequence of transactions is respected. Then, at the next transactions, the same validation process starts again, and again, and again.
Through this process, dependencies between the records are created, just like in a chain. That’s why this technology is called blockchain. If anyone wants to change the content of a block, they need to change its key and therefore need to change the keys in all the following blocks. Generating a key for a block can be made difficult using cryptography, consequently making tampering with the ledger very difficult, or even impossible. This is how an open ledger works.
A blockchain ledger is distributed to get rid of the intermediary authority. With everyone having a copy of the ledger, we won’t need the intermediary to store it. In a distributed ledger, whenever someone wants to add a record they would need to announce it to the full network and then all copies of the ledger are then updated accordingly. In action, to validate a transaction, we will need to make the network members compete to solve a difficult random mathematical problem that requires a lot of computation power.
For each block, the winner of the competition validates the transactions, signs the block, and adds it to the chain. In this construct, if a single person wants to insert fraudulent transactions in the blockchain, he needs to have more computing power than the rest of the network combined.
This is why blockchain, as a distributed open ledger, got a lot of traction. Not simply because it stores information, but because it creates trust without any third-party intermediation.
Underlying trends for Blockchain: Bitcoin and disintermediation
The fact that blockchain creates trust without the need for any third-party intermediation, gave it a natural start in the financial industry — a system that relies on a lot of trust in its institutions. In 2008, Satoshi Nakamoto, not his real name, wrote a paper describing a protocol to transfer digital cash between individuals without the need for bank intermediation and called it Bitcoin.
Bitcoin generated a lot of interest from investors in the overall technology and investment into Blockchain startups peaked in 2018. The volatility of Bitcoin as a cryptocurrency and few other incidents that happened over time, brought the investment amount down, over the years. But it also prompted tech entrepreneurs to build new blockchain or blockchain-based applications with the objective to completely dis intermediate established players.
The business value of Blockchain — disintermediation
In the case of buying a house, if all records of all houses in your city were in a secured trusted publicly available blockchain, then the title company would probably be replaced by an app that can check the ledger in a user-friendly manner for a minimal cost per transaction, or even for free if it subsidizes the service with a couple of advertising banners. Cutting out the middleman is not only saving users money, but it is fundamentally changing where business value can be created.
Bitcoin is probably the most popular example of blockchain disintermediation. Bitcoin relies on thousands and thousands of computers storing and exchanging replicas of an open ledger, using a specific blockchain protocol. To manage bitcoins in a user-friendly manner, there are digital wallets that manage interactions with the ledger, recording and storing transactions. The process of transferring bitcoins is a matter of adding a transaction in the open ledger. For such a transaction, a traditional bank will charge customers ~5% of the transaction amount, and the transfer would take a few days at best to complete. However, with bitcoin, transactions happen freely — at least for now -, and it will maybe take a couple of hours to be effective.
Due to bitcoins advantages, banks have started their private blockchain ventures. More than 70 banks around the world have founded the R3 consortium, to develop what they call a new operating system for financial markets, utilizing their private distributed ledger platform called Corda. This is a typical move from incumbents to protect their competitive advantage by adopting the technology or even reshaping it.
In addition to financial services, blockchain can disrupt and transform our democracies, by securely storing voting results. Whether it be a vote to elect a president, or a vote to clean the house. With blockchain, we can safely and securely track goods and services across their value chains, from raw materials to end products that we buy in retail stores.
Blockchain can also disintermediate certain and current platform-based businesses in what we might think of as the next wave of digitally disruptive companies. For example, in the case of Airbnb, the digital disruptor of the hotel industry, imagine that instead of advertising a property on a platform that takes a cut of the rents, the lesser publishes it on an Ethereum open ledger, with price tags and time windows. Ethereum is a blockchain that is built specifically to allow the creation of smart contracts — little pieces of code that can execute an action on the blockchain if certain events happen. So in action, one publishes property on Ethereum, and if another member of the network is interested, they set up a smart contract that will pay the agreed amount, take the deposit, and release it once the house is returned in perfect condition, all automated and no need for intermediaries, even Airbnb.
Challenges of blockchain: peripheral trust and operations
Blockchain generates trust only in the transactions that happened within the network, it will need another way to guarantee what is called the peripheral trust, or trust at the edge of the network. This can be some form of identity proof to guarantee that you are who you say you are or that the house represented in the network has a physical presence in the real world.
In general, the bigger the network, the cheaper it will be to transact, however, the more expensive it will be to guarantee trust in every single network node. Taking transaction costs and trust into account, an optimal size of a blockchain network might land somewhere in the middle. The specifics of the model will be defined by how blockchain adapts to a variety of unsolved questions such as, how can blockchain services be monetized?, who will capture most of its value more, the application layer or the protocol layer?, what should be the regulatory framework around blockchain?, or how to handle the capacity limitations of 10 to 30 transactions per second with a typical credit card system vs. handling 2500 transactions per second.
The rise of AI in the business world is powered by the combination of growing processing speed, learning algorithms, and availability of data. AI is currently progressing in the Narrow or Weak segment, and we are far away from reaching working General AI. AI market size is expected to reach ~$50 billion in 2020. Early adopters of AI technologies that can upscale their talent pool and enforce data governance best practices will keep a competitive edge.
Blockchain is a protocol to implement a distributed open ledger that securely keeps record without the need of a trusted third party. Although block chain is at the beginning of it’s journey, it has already applications in financial services, assets traceability, contract management, and many other domains. There are still multiple challenges associated with blockchain from technology difficulties to questions about the business model and the regulatory framework. However, if blockchain makes its way, it can completely disintermediate established platform players in today’s business ecosystem.
