Insuranceciooutlook

Can AI Help Blockchain Adoption?

By Aarthi Srinivasan, Director, Target

Aarthi Srinivasan, Director, Target

Funding has never been easier with $4B raised by Block. One in less than a year and $33M obtained in 30 seconds by SingularityNet. Over $11B in investments transpired in 2018 with 80 percent year over year growth in just the first half of this year.  Such is the growth trajectory of Initial Coin Offering (or ICO) cryptocurrency world.

"Blockchain can support a robust verification process during the data collection phase in these AI applications providing highly coveted data and feature set to the AI engine"

In comparison, the entire 2017 VC investments were about $70B (down by 4 percent YoY) and Uber took nearly 9 years to raise over $20B. If this ICO-mania (check out TataTu & EOS valuation) is making you anxious, you are not alone and regulators are working to clarify token rules. Irrespective of crypto prospects, blockchain technology (which became popular due to Satoshi Nakamoto's Oct 2008 Bitcoin paper) will impact data integrity and transparency. Blockchain will change the way companies think about Data collection, privacy, and analysis which are critical components of a robust Business Intelligence Strategy.

Fig1 – ICO annual growth as of May 2018

So, what is a blockchain?

1Assume we have a group of self-selected users (in permissionless blockchain where anyone can participate) or pre-selected invite-only users with specific roles (permissioned blockchain with private access) who execute blockchain software on their special computers (a.k.a nodes). This group of nodes constitutes the blockchain network.

2. Now let's imagine a constantly flowing supply of empty cardboard boxes representing blocks (1 MB block size for Bitcoin network.)

Fig 2 – A block

3. Users who want to store their transactions permanently in this box send (broadcast) their transactions such as financial transfers, barcodes, etc. to all (or some of the pre-selected in permissioned blockchain) the nodes in this blockchain network. The blockchain network will verify each transaction before it gets into the box. The software checks for a valid identity and the existence of sufficient account balance to transfer money.

Fig3 – A verified transaction

4Once the box is full with verified transactions (when the 1MB limit is reached), it is time to close the box, name, and seal it.

Fig4 – A filled block with verified transactions

5. The naming rights are granted to the person who finds the name first by solving a mathematical puzzle to produce a hashcode (alphanumeric string) with a leading number of zeros. This process of naming is called MINING and the winning person (miner) solves the puzzle first will be rewarded. In the Bitcoin network, the mining reward is currently set at 12.5 bitcoins and it will be halved after every 210,000 blocks.

Fig5 – A completed and named block with a hashcode

The formal definition for blockchain is "A digital ledger in which transactions are validated by a decentralized network of nodes (computers), chain linked and recorded chronologically."

Fig6 – A simplified blockchain prop

Having described the blockchain let us examine why we need it.

a) Reduces Cost by eliminating middlemen: If you want to transfer money directly between users you can use blockchain and eliminate the third party entities such as banks or card companies. This concept was implemented in the Syrian refugee camp in Jordan where refugee identities were captured in a blockchain along with each person's allocated amount of United Nation (UN) issued cryptocurrency. The refugees shopped in camp stores with their crypto balances and hence UN avoided huge transactional fees to 3rd party banks for every transaction.

b) Difficult to hack: Since the majority of nodes in the network have knowledge of all the verified transactions in the blockchain, it is not easy to modify transactions as the data discrepancy between nodes will become apparent if the information is not the same in the majority of the nodes.

c) Multi-user verification: The transactions in the block are verified by anonymous users other than the creator (centralized authority) of the transaction to avoid owner bias.  It will lead to fewer errors and lesser fraud as multiple verifiers are observing each transaction when it enters the blockchain. Verification can be done using a consensus method with few nodes approving transactions (less compute intensive) or using proof of work (more compute intensive) where a majority of nodes verify each transaction.

d) Append mode: If you need to edit a verified transaction in an earlier block, you cannot alter the block without modifying every verified transaction that was added later. Therefore, you have to add a new transaction with updates as opposed to modifying the earlier transaction in the block directly.

While Blockchain is already used in areas such as Finance, Payments, and Supply chain provenance, its growth can also be influenced by the technology investments in Artificial Intelligence.  In 2017, Artificial Intelligence (AI) investments grew by 141 percent YoY to ~$15.4B. US was the leading AI investor in summer of 2017 with 75 percent of equity funding share in AI companies. By the end of 2017, China was the AI investment leader with 48 percent equity funding share with the US following at 38 percent. The Chinese government supports investments in deep learning, vision chips as well as security using video surveillance& visual recognition. With the controversial "sharp eye" project, the government is experimenting to calculate a person's trustworthiness with a social credit score by monitoring video surveillance and facial recognition in various situations. Startup Megvii Face++ already has access to 1.3 billion face data records on Chinese citizens.

Additionally, AI investments in the sectors listed below will fuel blockchain adoption as a trustworthy data collection option.

a. Automobile technology (includes Autonomous Vehicles AV)
b. Facial Recognition & Security
c. Healthcare diagnosis
d. IoT data

Fig7 – AI growth areas

 

 

 

 

 

 

 

 

 

Blockchain can support a robust verification process during the data collection phase in these AI applications providing highly coveted data and feature set to the AI engine. This verification phase may encourage consumers or enterprises to share data pseudonymously in the blockchain to benefit from the reduced cost of transaction or safety. For example, if a blockchain is used to capture an Autonomous Vehicle's check-in/checkout locations & times at every stop with a consensus protocol certifying the entry, we can be confident that the data is verified by more just a single centralized AV's controller and maintained in an immutable fashion. Now you can confidently execute a route prediction or exception analysis.

So, what?

In the next five to seven years, blockchain will be a viable option for trustworthy data collection, identity platforms and payments. In 10+ years, AI applications will start making predictive decisions on verified data to execute smart contracts with explicit approval from us potentially using our blockchain enabled smart devices!

I hope this article has inspired you to start exploring blockchain ideas where data integrity is paramount to solving a real customer problem - Dream big!

Read Also

AI-Induced Technological Shift in the Insurance Industry

AI-Induced Technological Shift in the Insurance Industry

Abel Travis, VP, Underwriting and Product Innovation, AF Group
Analyze This: How Fresh Takes on Data Analytics are Driving Business Growth

Analyze This: How Fresh Takes on Data Analytics are Driving Business Growth

Doug Hague, EVP & Chief Analytics Officer, Bank of America Merchant Services [NYSE:BAC]
Data Analytics for Connected Cars

Data Analytics for Connected Cars

Hillol Kargupta, President, Agnik
Balancing Priorities in a Transactional Environment

Balancing Priorities in a Transactional Environment

Behz Naini, VP, Giesecke & Devrient Systems Canada, Inc.

Weekly Brief

Top 10 Artificial Intelligence Solution Companies - 2019

Artificial Intelligence Special