DEV Community

Cover image for AI-ML-Data Governance & Controls: Why ChatGPT-LLMs-Generative AI Cannot Be Trusted: Why We Still Need to Advance R&D on Them
Dr. Yogesh Malhotra
Dr. Yogesh Malhotra

Posted on

AI-ML-Data Governance & Controls: Why ChatGPT-LLMs-Generative AI Cannot Be Trusted: Why We Still Need to Advance R&D on Them

2023 New York State Cybersecurity Conference, 25th Anniversary Cybersecurity Conference, New York State Capitol, Albany, New York, June 6-7, 2023 ( https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4444658 : Originally Published on SSRN: 2023 New York State Cybersecurity Conference invited presentation: Sequel to: 2022 New York State Cyber Security Conference: Invited Presentation:
How You Can Implement Well-Architected ‘Zero Trust’ Hybrid-Cloud Computing Beyond ‘Lift and Shift’: Cloud-Enabled Digital Innovation at Scale with Infrastructure as Code (IaC), DevSecOps and MLops : https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4131044 : YouTube Video Presentation: https://youtu.be/1gd6mKeHJuw).

Post-Script: To bring much needed objectivity to the current global conversations about AI, we provide scientific and empirical evidence to put to rest many of the extreme myths, both utopian and dystopian, to share the much needed balanced understanding of AI as published in our latest research journal articles as well as presented at the recent four New York State Conferences on Cybersecurity sponsored by the State Governor.

The specific much needed progress to catch up with today's empirical realities necessitates AI and Data Science in all their flavors to catch up with the following original R&D leading practice progressions that we have delineated in our four recent New York State presentations.

AWS-Quantum Valley Building Quantum Minds For Quantum Uncertainty
• AI Automation -> AI Augmentation-Model Risk Management
• AI Cyber Controls -> Counter-Adversarial Command-Control
• Data Management -> Quantum Uncertainty-Complexity
• Data-Driven Architectures -> Event-Driven Architectures
• AI Automation -> AI Augmentation-Model Risk Management

AI-Machine Learning Augmentation and Cybersecurity: Why Smart Minds Using Smart Tools Are Critical for Minimizing Risks, And, What You Can Do About It? 2019 New York State Cyber Security Conference, Albany, NY, June 4 - 5, 2019, Empire State Plaza , Albany, NY

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3399781

• AI Cyber Controls -> Counter-Adversarial Command-Control

C4I-Cyber Command & Control Supremacy: Why It’s More Critical Than AI & Quantum Supremacy & What You Can Do about It? Security in Post-COVID Virtual Era Beyond Data, Models, Algorithms. 2021 New York State Cyber Security Conference, June 8-9, 2021, Empire State Plaza - Albany, NY.

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3851807

• Data Management -> Quantum Uncertainty-Complexity

How You Can Implement Well-Architected ‘Zero Trust’ Hybrid-Cloud Computing Beyond ‘Lift and Shift’: Cloud-Enabled Digital Innovation at Scale with Infrastructure as Code (IaC), DevSecOps and MLops, 2022 New York State Cyber Security Conference: Invited Presentations, Albany, New York: https://its.ny.gov/2022-nyscsc.

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4131044

• Data-Driven Architectures -> Event-Driven Architectures

AI-Machine Learning-Data Management Governance & Controls Cloud Computing Best Practices: Why ChatGPT-LLMs-Generative AI Cannot Be Trusted: Why We Still Need to Advance R&D on Them: Beyond AI Hype: Advancing Beyond Limitations of ChatGPT, Large Language Models, and Generative AI: CNY Quantum Valley Pentagon-USAF-USSF Ventures Spanning Air-Space-Cyberspace-Outer Space. 2023 New York State Cybersecurity Conference, 25th Anniversary Cybersecurity Conference, New York State Capitol, Albany, New York, June 6-7, 2023.

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4444658

Pragmatic Necessity of Bringing Scientific Objectivity to Current AI Debates

AWS-Quantum Valley Building the Future of AI-Quantum Networks: Next 30-Years Beyond Past 30: 4 years before today's 'fashionable' worldwide conversations about Existential Risks posed by AI, we had not already built worldwide recognition of such risks but also built the US Air Force-Air Force Research Lab Ventures AIMLExchange.com and C4I-Cyber.com specifically focused on controlling and mitigating such risks. The latest AI-Science headline discussions are still playing catch-up to where we started about thirty years ago, particularly in the context of Risk Management. Specifically, the latest ISO 31000 Risk Management Standard is yet to catch-up with our last three-decades related AI-Cyber-Risk Management R&D leading worldwide practices on Risk, Uncertainty and Complexity.

The latest global AI hype on Generative AI, Large Language Models, and ChatGPT has even thrown the standard practices of Data Cleaning for Data Integrity, Validity and Reliability out the window by unquestioning reliance on underlying databases such as Wikipedia that are known to have no scientific objectivity and reliability given over-predominant absence of scientific review and refereeing processes that are characteristic of the world-class journal and conference proceedings, as one who has had the privilege of being invited to serve on more than 50 such global top-tier world class STEM research journals and conference proceeding realizes.

The latest Generative AI hype has sacrificed Data Integrity and Reliability and degraded AI to a scientific reliability and validity standard lower than its prior focus on Data Science that had already been struggling to sustain its credibility by shuffling around from Big Data to Small Data to Right Data. Even the prior focus of Data Science with its backward looking focus on Historical Data being used for Prediction of the Future had already made all related research findings questionable as Future is not equal to Past. Given these observations, our R&D published in world's top refereed STEM research journals as well as invited service as STEM editor and referee for world's most esteemed research journals and conferences were founded on departing away from focus on [Historical] Data Management to Managing Emerging Uncertainty and Complexity for Future Business Enterprise Performance in early 1990s. The application of our three-decade R&D in our own global practices leading world's largest governments and enterprises as well as its empirical validation in worldwide adoption by hundreds of other world's largest governments and enterprises lends credence to our scientific R&D.

This fact becomes all the more critical for the latest era of most highly unprecedented Uncertainty and Complexity where world leaders from those leading hi-tech nations to world's largest governments and organizations are attempting to make sense of the latest AI hype. Our latest 2023 New York State Cybersecurity Conference focus in the latest presentation brings the much needed clarity while explaining the deviant human behaviors underlying the latest unscientific Generative AI-hype contradicting established scientific evidence. Specific evidence that we share includes scientific facts about the human-machine systems and relative resilience of humans vs. machines in navigating uncertainty such as based on the actual scientific comparisons of human brains and artificial intelligence machines as well as the foundational science of entropy of human-machine systems that has been overwhelmed by the world's centi-billionaires most highly questionable headline making claims about the untested superiority of artificial intelligence machines particularly in the face of overwhelming real world empirical evidence such as the hyped Self-Driving cars on which more than $100 Billion has been squandered given reliance upon hype pumping the financial stock markets which is now widely self-evident in contrast to rigorous scientific research foundations.

ABSTRACT

Network-Centric Meaning-Driven Human-Centric AI-Cyber Computing Beyond Data-Driven to Event-Driven Architectures for Quantum Uncertainty, 1995-2023:

Building upon the contextual focus of current global worldwide discussions on GPT, ChatGPT, GenAI, Generative AI, Large Language Model - LLMs, we help you advance beyond the ongoing global AI-hype - on both dystopian and utopian extremes - to focus on the latest AI Event-Driven Architectures technical developments such as in AIOps, MLOPs, DevSecOps, Infrastructure As Code, Configuration As Code, Platform As Code, Pipeline As Code for building Cloud Computing AI-Agility and Cyber-Resilience Sustainability, the focus of our last year's 2022 New York State Cybersecurity Conference presentation as AWS Partner. With increasing Digitization of Networks as Code, we help world's global 'Hardware', 'Software' and other 'Computing' providers and practitioners advance beyond legacy models of computing to the latest Cloud-based Networked Computing Utility models. Our 30-year R&D focus on Network-Centric Computing from both Socio-technical and Systems Engineering perspectives spans the schisms presented by Claude Shannon's Information Theory in how AI-ML enabled Information Processing and Sense Making can occur across both Socio-Technical and Computing-Telecom Networks*.

Having already addressed the core issues being debated on mainstream Business, Finance, Technology and other media and resolved them based on 30-years of our AI-Cyber-Network Science-Engineering R&D leading worldwide Digital practices, we also help you advance ahead by 25-years on the global Risk Management standards given that the latest ISO 31000 Risk Management Standard is playing catch up having lagged 25-years behind our R&D leading global Risk Management practices. Based on most thorough R&D over 30-years and most in-depth analysis of current academic, scholarly, policy and practice resources, we provide definitive answers on many issues that have left too many confused with what our analysis shows as extreme dystopian and utopian views. This confusion results from ambiguity between the specific AI technologies and their use, abuse, misuse by the respective human users in diverse contexts, our core R&D focus on Human-Machine Systems Learning, Intelligence and Performance for 30-years ranked for its worldwide impact among AI-Quant Finance Nobel Laureates such as Herbert Simon (ASIS&T and University of Minnesota research impact and citation impact reports, for example) and Black-Scholes, Markowitz, and Sharpe (AACSB Impact of Research Report, for example) with our related research models and methods applied in empirical practice by organizations as diverse as the National Aeronautics and Space Administration (NASA) and Big Banks, for instance.

• Should we be afraid of latest AI such as Generative AI and ban such R&D?

  • No! Such R&D must continue by dispelling the myths being spun about AI.

• Does AI have the capability or capacity of posing existential risk to humans?

  • No! But Humans with vested interest in exploiting other Humans do!

• Does AI have intelligence to take over control of humans & destroy them?

  • No! But Humans with vested interest in exploiting other Humans do!

• What can be done for dispelling the AI myths spun up to make real progress?

  • We have already been doing so having already advanced beyond such myths with focus on saving all 90% time, cost, & resources by our focus on Science & Scientific R&D driven Zero-Hype practices.
  • We share our learning and freely accessible R&D leading global practices with all in order to help you advance beyond the AI-Hype in making real progress for the human civilization and humanity at large.

  • There has been an over-concentration on Shannon's definition of information in terms of uncertainty (a very good definition for the original purposes) with little attempt to understand how MEANING directs a message in a network. This, combined with a concentration on end-points (equilibria) rather than properties of the trajectory (move sequence) in games has led to a very unsatisfactory treatment of the dynamics of organizations. – AI-Genetic Algorithms pioneer Dr. John H. Holland, then at the Santa Fe Institute (personal communication, June 21, 1995) :

Source: Malhotra, Y., Expert Systems for Knowledge Management: Crossing the Chasm between Information Processing and Sense Making, Expert Systems with Applications: An International Journal, 20(1), 7-16, 2001. (Ranked by the journal as top-ranked article in its usage statistics).

Keywords: Artificial Intelligence, Machine Learning, Data-Driven Architecture, Event-Driven Architecture, Quantum Uncertainty, Network-Centric Computing, Networks as Code, Cloud Networks Engineering, Cloud Cybersecurity Operations, AIOps, MLOPs, DevSecOps, Infrastructure As Code, Configuration As Code

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Image description

Top comments (0)