Michael Proksch, PhD

Without skilled data engineers, D&A initiatives face higher costs, delays, and data quality issues. [link] Leaders must act now—upskill teams and hire for the future of cloud and GenAI. Discover essential skills guide to identify and develop skills for data engineering: https://gtnr.it/47NR8qc #GartnerIT #Data #Analytics #AI #Cloud

40

Advancing AI Investigative Interviewing: A Grant Fraud Case Study Model Empowered by Executive Order 14179 By Professor Gary Vause, II CEO, Vause Computer Systems, LLC | Faculty, Institute of Artificial Intelligence and Robotics Artificial intelligence is fundamentally reshaping the landscape of investigative interviewing and fraud detection. In my latest research, I introduce the Grant Fraud Investigation Case Study Model, which integrates AI-driven cognitive interviewing tools to help agencies assess truthfulness, validate evidence, and fight complex fraud schemes in real time. This work would not be possible without the continued partnership and support of the Department of Homeland Security. DHS’s collaborative efforts have enabled us to test, refine, and implement AI-powered investigative tools that meet the urgent needs of federal and local agencies. Our work is directly aligned with Executive Order 14179, issued on January 23, 2025. This order removes ideological and regulatory barriers to AI innovation and calls for the development of AI systems that are free from bias. Its overarching goal is to secure America’s leadership in artificial intelligence while advancing national security, economic competitiveness, and public trust. At Vause Computer Systems, LLC, we take this mandate seriously. As a Black Veteran-owned technology management consulting firm established in 1990, we are proud to be a trusted advisor to public and private sector clients. We are leading the way in applying AI for truth verification, fraud analytics, cyber governance, and intelligent risk detection. We see promise in technologies like digital polygraphs, identified among the top strategic tech trends of 2025, which are shaping the future of disinformation security. These tools—paired with psychological interviewing techniques such as the Cognitive Interview, Strategic Use of Evidence, and Criteria-Based Content Analysis—are creating new possibilities for hybrid investigative intelligence. Our case study model is more than academic research. It is a working prototype for how federal agencies can use science-based AI technologies to interrogate evidence ethically, identify deception reliably, and build stronger fraud cases. To DHS, thank you for being an invaluable partner. To those in law enforcement, federal investigations, and digital forensics, we invite you to join us in shaping a new era of AI-assisted truth assessment. Let’s build systems that protect the innocent, expose the truth, and uphold the highest standards of justice in the digital age. Gary Vause, II Founder and CEO, Vause Computer Systems, LLC LinkedIn: https://lnkd.in/ebUAxS_X

7

Happy to share these resources on infrastructure technical debt that are available to everyone! 40% of infrastructure systems across asset classes have technical debt concerns: https://gtnr.it/3FGoThY Without targeting the main issues, temporary modernization efforts won't make a real difference. These steps will. Gartner for IT| #GartnerIT #TechnicalDebt #Assets #Enterprise

40

1 Comment

So... I am very buttoned up about citations, but the fake citations have now hit me personally. This past week, I went over edits I got back on a conference proceeding. It had comments and redlines done with MS Word track changes. Welp, I was looking at the text before the redlines, and some things looked sloppy. I didn't think I would write like that. In MS Word, I did reject all tracked changes, and then I diffed what I had gotten back from the editorial board against my original submission. The diff was lit up. There were numerous changes that had been made without track changes. In the body of the article, it was mostly things like spelling out a number vs using the numeral, and what I assume is house style. I don't *think* the meaning of anything changed, although it was A LOT and I wish they had given a heads up so I could review. When I compared citations, it was a complete mess to the point of being research misconduct. An editor had removed citations to my sources which I consulted and which supported the parts of the article on which each appeared, and had put in citations to different sources which I didn't consult. About 2/3 of the citations no longer supported the text of the article, and at least one appears to be a citation to a non-existent source. And I was never informed. The editor did that before turning on track changes. I don't even know which of the three editors put the fake cites in, because of the person never informing me nor recorded any kind of notes on the article. I've put my correct cites back in, and I will email it back. I have misgivings because for all I know, this same editor could do this later in copy editing. Alls I can think is everyone out there needs to learn how to diff, and to make it part of the process: Each draft you get back, diff your original against the draft you got back, review the unmarked changes, and check that the editors did not insert fabricated sources. And, pretty much, I now have to make this part of the process, to where each time I get a draft back, I need to check that the editor did not insert fabricated sources.

123

23 Comments

An important article I wrote that differentiates LLM assistance and agentic AI. Please read. 🔗 https://lnkd.in/eb8mJyjJ A note for those applying to jobs: This matters for your career too. When you're in an interview and someone asks what you built, be precise. If you used an LLM to help you write code faster, say that, don't call it an agent. If you built a RAG pipeline, call it retrieval-augmented generation, don't call it agentic. If you built a chatbot with a system prompt, own it for what it is. Hiring managers who know the difference will respect your precision. Hiring managers who don't know the difference need you to teach them. Either way, you win by naming it correctly. Overselling what you built doesn't just hurt your credibility when someone digs deeper, it tells the room you don't actually understand the architecture you're describing. Know what you built. Know what it's called. Know why it matters at that level. That's what gets you hired. #AI #AgenticAI #LLM #NLP #SportsAnalytics #DataScience #CareerAdvice #MachineLearning

13