Author: Yogi Nelson

Hello, I'm Yogi Nelson. Reading, writing, and learning have always been my passions. For 33 years I worked in the public sector in the fields of city planning, housing, real estate, and finance. I retired in 2017 and travel extensively through Asia while dedicating myself to yoga. In 2020, I decided to dive into blockchain and crypto and now artificial intelligence. Join me as I answer your questions on these topics. As a bonus, I include proverbs from my upcoming book, "Global Wisdom: Proverbs from Every Nation", nano size comments related to one of my travel adventures and fun facts from different countries.
AI Agents, Artificial Intelligence, Blockchains, Environment, Productivity, Railwlays, Uncategorized, Urban Planning, Yogi Nelson

Navigating AI Disruption in the Railway Sector

Yogi Nelson

Welcome to the Blockchain & AI Forum, where your technology questions are answered.   Harold asks, how will artificial intelligence affect the railway industry?

Harold, you came to right place.  I’ll answer your question by citing research reports from McKinsey & Company, International Union of Railways (UIC), and Startus Insights.  Let’s start with the McKinsey report, titled: The Journey Toward AI-Enabled Railway Companies.

Prominent AI Use Cases in the Railway Industry:  McKinsey Report Findings

According to McKinsey, railway companies are focusing their AI efforts on approximately 20 use cases.  It should be noted, the McKinsey identified over 100 potential use cases!  The top 20 use cases are related to four key performance indicators:  on-time performance, customer engagement, safety, and operational performance.  The table below summarizes the use cases by category:

Table 1
ARTIFICIAL INTELLIGENCE USE CASES IN THE RAILWAY INDUSTRY
Business ActivityDescriptionAI Use Case
Railway UndertakingsActivities that are directly & indirectly related to delivering the final productCrew trainingSafetyOnboard operationsRolling stock maintenanceTrain station operationsRolling stock procurement & inventory management
Infrastructure ManagementActivities that are directly & indirectly related to the infrastructure needed to deliver the final railway productNetwork planningNetwork traffic managementNetwork constructionNetwork maintenance Rail safetyNetwork procurement and inventory management
Passenger ExperienceCustomer facing activities related to the operations of the railway companyBooking and ticketingCustomer railway journeyPassenger informationIn-station additional revenue  
Support FunctionsEssential non-operational activitiesHRFinanceIT

Of course, not all use cases have equal priority.  In the table below, the five most common use cases are listed on the left, and the five least used on the right.  The other 10 are in the middle.

Table 2 ATable 2B
Top 5 AI Use Case in Railway IndustryBottom 5 AI Use Cases in Railway Industry
Crew and shift managementNetwork infrastructure digital twin
Rail predictive maintenanceReal time disruption management
Real time intermodal informationTalent training
Energy efficient managementAutonomous trains
Security fraudSoftware development

The Keys to Successful Incorporation of AI in the Railway Industry:  UIC Report

We now turn to the findings by the Union of International Railways; UIC per their French acronym.  The UIC researchers claims that the railways with the most successful AI abilities share certain characteristics. For example, these railway companies dedicate research and development teams to their AI efforts.  Furthermore, they establish cultures of innovation and investment in partnerships to develop new technology.  Yet a third reason is that they build capabilities to implement use cases; and take a business-driven approach (rather than solely relying on IT departments) to drive development.  UIC researches say, there are six specific keys to success: strategic roadmap, skills, agile operating model, technology, data, and adoption and scaling. 

Who Are the AI Disrupters in the Railway Industry:  StartUs Insights Report

According to StartUs Insights, a consultancy specializes in discovery of intelligence trends, innovation, and engagement with the world’s most promising startups, scaleups, and technologies, there are 20 AI startups impacting the railway industry.  I’ll refrain from listing them all, and instead highlight three that best exemplify the AI story in railways. 

Let’s start with rail safety and the startup called RMT.  RMT is based in Italy.  RMT specializes in automatic recognition of defects on rail infrastructure, and employing AI to detect physical defects.  Energy efficiency is the wheelhouse of RailVision Analytics, a Canadian startup dedicated to improving energy efficiency via algorithms that reduce unnecessary braking, throttles, and train handling adjustments.  Last, there is Cervello.  Cervello is an Israeli startup. Cervello says it provides railways with “… comprehensive security solutions and in-depth visibility of the operational network and infrastructure.” With Cervello misleading commands are identified.  Moreover, Cervello produces a contextual view of the affected assets and operational impact.

Time to go but not before ending with this famous lyric from the O’Jays:  …” People from all over the world join hands start a love train, love train …”

Until next time,

Yogi Nelson

Uncategorized

THE USE OF ARTIFICIAL INTELLIGENCE IN HEALTH CARE:  A LITERATURE REVIEW

Yogi Nelson

Welcome to the Blockchain & AI Forum, where your technology questions are answered! Today’s question, from Doctor Cristina Lazaro, concerns the use of artificial intelligence in health care.

Doctor Lazaro, you came to the right place.  There has been an explosion in research related to AI in health care. I answer your question based on an article titled, Artificial Intelligence in Healthcare Institutions: A Systematic Literature Review on Influencing Factors by Roppelt, Kanbach, and Kraus, 2024 , Science Direct.  https://www.sciencedirect.com/science/article/pii/S0160791X23002488

The Health Care Status Quo–Serious Challenges Ahead with Ai as a Solution. According to researchers, the global health system is facing serious challenges.  Researchers cite the growing number of patients due to population growth, increased prevalence of chronic diseases, increasing healthcare costs, and staff shortages.   The good news is AI has the potential to solve some of the challenges ahead, claim the authors.  The researchers say, “… AI is a technology that, through rules-based logic, can help to significantly speed up the process of analyzing vast amounts of data and leverage patterns by mimicking human intelligence.  AI may result in fast and often better-advised decisions. Thus, as it pertains to the healthcare industry, AI has the potential to overcome staff shortages in developing and developed countries, enhance organizational efficiency, and maximize diagnostic accuracy as well as patient outcomes by providing at least comparable quality results compared to human based assessments. Consequently, Al may reduce costs due to avoidance of inefficiencies, unnecessary treatments, and late diagnoses.

Literature Review. Researchers discovered approximately 1,100 published articles on the topic since 2023.  However, after analyzing the literature in greater detailed soon learned about 130 were truly on point.  The authors also found that most studies were published in journals related to information technology, not medicine, or journals dedicated to the nexus of AI and medicine.  AI seems to be most widely adopted in cardiology and radiology and rather at an early stage in dermatology and psychiatry.  Interesting that there are not many studies of AI in medicine within business management journals.

FACTORS THAT INFLUENCE AI ADOPTION IN HEALTHCARE:  THE FINDINGS

Macro-economic readiness. Researchers at Science Direct found that macro-economic readiness is mainly driven by governments.  For example, researchers conclude adequate IT infrastructure, e.g., access to smart devices for the population and energy adequacy, especially for remote areas and certain populations drive adoption.  “AI communities”, is also a critical.  Essentially, when there are established partnerships, collaborations, etc. the network effects kicks in and the impact is exponential. 

Technological readiness.  Researchers concluded that technological readiness depends on three factors: 

  • The importance of overcoming algorithmic challenges, e.g. lack of data, and/or inferior quality, etc.
  • Provide a compelling, multi-faceted value proposition.  In other words, clearly added value in the form of efficiency, user-friendliness, adherence to data-privacy regulations, interoperability, and the ability to tailor the application to individual needs.

Regulatory readiness.  The authors suggest close collaborations with regulators.  The author points to three key issues:  political support, clarification of legal questions, and establishment of regulations. As a former government regulator, this is sage advice but not easy to execute because contrary to popular belief, the “government” is not a unified organism.

Organizational readiness.  A failure to plan is a plan to failure, say researchers, hence they strongly recommend preparing the organization before implementing an AI health care system.  Researchers recommend four steps:  prepare an organizational strategy, develop a supportive organizational culture, clear tasks and assignments, and an adequate IT setup.

User readiness. Recognizing that organizations consists of people, researchers says place users are at the center.  Furthermore, researchers conclude four facts are the key to user readiness:

  • Awareness:  in other words, do users know the technology exist?
  • Beliefs:  do users believe the technology will work? 
  • Personal innovativeness:  are users willing to try something new?
  • The financial situation of the individual.

Doctor Lazaro, now you know what the literature says. I end with a proverb from Spain: ”donde hay hambre, no hay pan duro.”

Until Next Time,

Yogi Nelson

Uncategorized

Tornado Cash Legal Update: OFAC Sanctions Lifted

Yogi Nelson

Welcome to the Blockchain & AI Forum, where your questions are answered!  Today I answer Malik’s question: “what’s the latest news in the Tornado Cash legal case”?

What is Tornado Cash. Before jumping to the answer, first a moment to understand what is Tornado Cash.  Tornado Cash is a non-custodialprivacy protocol allowing permissionless shielded transactions on Ethereum and derivative networks. In plain English, Tornado Cash is an application that does not hold your crypto currency–you own it, you store it.  Tornado Cash facilitates transactions on the Ethereum blockchain without requiring permission while maintaining your privacy.  How does Tornado Cash maintains privacy?

Tornado Cash = Privacy.  Tornado Cash achieves anonymity by breaking the on-chain link between source and destination addresses through “anonymity pools”.  However, Tornado Cash does not preserve privacy by comingling assets; it uses zero knowledge cryptography.  What’s more, Tornado Cash relies on Ethereum—a decentralized protocol where transactions neither can be changed or altered.  Hence, nobody – including the original developers – can modify or inhibit the protocol or the transactions. What more information about how Tornado Cash works?  Click   How does Tornado Cash work? and Tips to remain anonymous.

What is the Office of Foreign Assets Control—OFAC.  OFAC is an office within the U.S. Treasury Department.  It claims to enforce economic and trade sanctions against specific countries, regimes, individuals, and entitles that pose threats to U.S. national security, foreign policy, or economy.  In other words, the enemy de jour of the U.S. government.  By the way, although OFAC was formally created in 1950, the use of sanctions by the U.S. government goes back to 1812! 

Tornado Cash Placed on OFAC Sanction List. August 8, 2022, OFAC placed Tornado Cash on the sanctions list.  According to the OFAC press release https://home.treasury.gov/news/press-releases/jy0916, “Tornado Cash, is a crypto currency mixer which has been used to launder more than $7 billion worth of virtual currency since its creation in 2019.”  However, OFAC’s statement is partially incorrect on the facts.  Tornado Cash is not a currency mixer nor is it property.  This error, deliberate or not, is materially important as we will see in a moment.  Let’s move on to what happened last week.

OFAC Removes Tornado Cash from the Sanction List.  Last week, the U.S. Treasury removed Tornado Cash from its sanctions list.  That’s a big deal. To understand what happened let’s take a moment to examine a recent court ruling that led to OFAC removing Tornado Cash from the sanctions list.

Last November, the 5th District Court of Appeals ruled that OFAC exceeded its authority when it put Tornado Cash on the sanctions list.  Essentially, the Court of Appeals reversed an earlier decision by a lower federal court.  The Court of Appeals stated the executive branch’s authority to “block ‘property’ in which a foreign ‘national’ or ‘person’ has an ‘interest’” did not apply in the case of Tornado Cash.  In other words, OFAC has no jurisdiction.  Why?  Because the use of immutable smart contracts — lines of autonomous code on the blockchain intended to preserve anonymity in transactions — do not qualify as property and OFAC can only sanction property.  And, smart contracts, despite their names, are not actually contracts. By the way, Tornado Cash’s legal fees were paid by Coinbase.  Wonderful to have friends with cash! 

No surprise—Tornado Cash value catapulted by 73% after the OFAC full scale retreat.  The story ends with OFAC deciding to “exercise its discretion” by dropping legal actions against Tornado Cash.  Lol! 

Until Next Time,

Yogi Nelson

Artificial Intelligence, Banking, Blockchains, cryptography, International Finance, Uncategorized

GENISUS Act: Implications for Stablecoin Issuers

Yogi Nelson

Namaste Yogis.   Welcome to the Blockchain & AI Forum, where your questions are answered!   As a bonus, a proverb is also included.  Today’s question comes from Latasha and she asks what is the GENISUS Act?

The GENISUS Act is a proposed federal law, introduced by Senator Hagerty of Tennessee–your home state.  To no surprise, GENISUS is an acronym: “Guiding and Establishing National Innovation for U.S. Stablecoins, 2025.” (Feds have the best acronyms!  Lol.)  Let’s review GENISUS Act section-by-section, beginning with definitions.

GENISUS defines 23 terms but none more important than Payment Stablecoin. According to GENISUS, a payment stablecoins is a digital asset that is designed to be used as a method of payment/settlement, and which the issuer is obligated to convert, redeem, or repurchased for a fixed amount of monetary value and represents it will maintain or creates the reasonable expectation that it will maintain, a stable value relative to the value of a fixed amount of monetary value, and is not a national currency or a security issued by an investment company.” We now move on to Sections 3 – 7.

Permitted Stablecoin Issuers:  If GENISUS becomes law only a permitted stablecoin issuer can issue stablecoins.  No free lancers permitted!  Don’t even consider it unless you enjoy living in a 6 x 9-foot prison cell with a dude name Buba!  Lol.

Requirements for Issuing Payment Stablecoins. There are a litany of requirements to issue a payment stablecoin.  Below are a few highlights of what issuers must do:

  1. Maintain reserves backing the issuer’s payment stablecoins outstanding on an at least 1 to 1 basis comprising of: U.S. Dollars, demand deposits held in U.S. banks, treasury bills, notes, and bonds of 93 days or less, repurchase agreements of seven days or less, money market funds, Central Bank reserve deposits.  In other words, liquid reserves that are 1-to-1. 
  2. Issuers must also publicly disclose their redemption policy, establish procedures for redemption, and publish monthly composition of the issuer’s reserve. Disclosure!
  3. Monthly compliance certificates from the issuers, CEO and CFO.

Limitations of Stablecoin Activities.  Stablecoin issuers are limited to the following activities:

  • Issue payment stablecoins
  • Redeem payment stablecoins
  • Manage related reserves
  • Provide custodial or safekeeping services for payment stablecoins
  • Other directly related work.

Stablecoin Regulators.  GENISUS does not establish a new federal stablecoin regulator.  Relief! Consequently, if a nationally chartered bank, e.g. Bank of America, issues a stablecoin the same federal banking regulator, Office of the Comptroller of the Currency, (OCC) that supervised Bank of America for safety and soundness, will be tasked with regulating the bank’s stablecoin operations.  Simple.  The same applies if the bank has a FDIC banking charter.  State chartered banks can issue stablecoins, provided their corresponding state banking regulators follows the GENISUS Act. Standard practice in banking.  What if a federal qualified nonbank wants to issue a stablecoin you ask?  OCC will be the regulator of those nonbank stablecoin issuers.  Let’s move on.

Application and Approvals.  From the date the stablecoin regulator deems the application complete, the applicant must be approved or denied within 120 days.  The catch is from the date the application is deemed complete.  A malicious regulator can string an applicant along with a constant request for additional information. 

If the regulator deems the application unsafe it can reject the application but it must explain with specificity, including all findings made by the regulator with respect to all identified material shortcomings in the application, including actionable recommendations on how the applicant could address the identified material shortcomings.  I find this last section unusual and troublesome.  Essentially GENISUS requires the regulator to explain to the applicant how to do their “homework”.  This is problematic.  Afterall, if an applicant is incapable of preparing a coherent application, doesn’t it portent an institution incapable of running the program effectively?  I think so.  Moreover, if the stablecoin issuer fails, will it blame the banking regulator for its failure given the issuer could reasonably claim it followed the regulator’s instruction? 

Rulemaking.  Did you know that when Congress passed a law it writes into the law a section titled rulemaking?  The purpose of this section is to authorized and require the executive branch to write the regulations necessary to implement the letter and spirit of the law.  In the case of GENISUS, Congress directed the various stablecoin regulators to issue the regulations within 180 days from the day GENISUS becomes law.  Regulators, you better step on the gas!

Time for me to make a genesis appearance at my next engagement but before I go, I leave you with this proverb from Turkey: “a hungry bear does not dance”.

Until next time,

Yogi Nelson

Artificial Intelligence, Blockchains, cryptography, Germany, Uncategorized

Latest Trends in Crypto Key Management: Insights from Germany

Yogi Nelson

Namaste Yogis.   Welcome to the Blockchain & AI Forum, where your blockchain technology questions are answered!  As a bonus, a proverb is also included.  Today’s question comes from Wolfgang in Stuttgart and he wants to know what is the latest in crypto currency key management from Germany?

Wolfgang, you came to the right place.  I understand why you would ask this question and its importance.  After all, as often said in the crypto world, “not your keys, not your crypto!”.  With that in mind I drove deep into academic research for the latest findings.  I discovered an article titled, “Perceptions of Distributed Ledger Technology Key Management – An Interview Study with Finance Professionals”, Guthoff, et al, 2023, Helmhotz Center for Information Security, Germany.  Their paper is focused on key management for financial institutions not individuals, nevertheless it is fascinating to learn what the BIG BOYS & GIRLS are thinking. .https://publications.cispa.saarland/3948/1/guthoff2023keymanagement.pdf

RESEARCH BACKGROUND AND METHODOLOGY:  The researchers focused their attention on private key management within the traditional financial sector (trad-fi) and the cryptocurrency space (de-fi).  By incorporating the views and perspectives of employees working in trad-fi and de-fi, the unique perspectives of the established and newer financial institutions were both included. Nice!  In section two, the authors define a number of essential terms readers must understand to completely comprehend the findings.  Below are two examples of terms that were defined:

Key Management:  the activities involving the handling of cryptographic keys and other related key information during the entire life cycle of the keys, including their generation, storage, establishment, entry and output, use, and destruction. According Guthoff, key management is a combination of different key-related aspects, including key usage and storage.  Guthoff explains further that the goal of key management is to generate, store, distribute, use, and revoke cryptographic keys (integrated into enterprise processes) while keeping the private key secret. Key Management is not necessarily equal to a (cryptocurrency) wallet, but a wallet may be used as a simple tool for key management.

Custodial Key Storage: is the storage of keys by a separate party, which can be either in-house or a third entity that does not use the key. This party can, but does not have to, take care of all other aspects of KM. For self-managed assets, this could be a cryptocurrency exchange. For financial institutions, custodial key storage can be implemented in-house or externally by an intermediary.

Research Findings:  Essentially, the threats to key management fall into four different categories: internal, external, systems and infrastructure, and force majeure.  Force majeure means unforeseeable circumstances.  Unfortunately, the research paper covers too many findings for me to discuss them all; hence, instead I’ll give you just a few highlights.

Multi-signature keys:  Multi-signature keys were a hot topic among respondent and participants were enthusiastic about the concept.  According to the survey participants the two main benefits of this technique were resilience against key loss or compromise and added protection against malicious employees. Some participants also proposed storage of keys at a bank or a notary as a backup option.

Invisible Key Usage. The invisible key usage pertains to trad-fi but has implications for de-because management of private keys definitely holds back mass adoption.  In the view of the survey participants, they preferred a solution where key usage is mostly encapsulated within an application and hidden away from the user.  Although participants were referring to enterprise systems, the principle applies–the risk of losing a key through an attack or intended malicious actions by an employee.  Participants thought usage of the key is best through access via personal credentials like username and password or multifactor authentication, using a phone or biometric features.

CONCLUSION:  Bad news–researchers are skeptical whether existing key management systems can function effectively at the enterprise scale.  They say efforts to maintain on-chain authentication and audit functionality covering day-to-day operations, e.g. employee absence or shifting responsibilities within the organization – as well as – contingencies such as key compromise or force majeure events – is unmanageable, except for the smallest institutions. 

Time to say goodbye with a proverb from Germany: “one becomes wise through damage”.

Until next time,

Yogi Nelson