1. RBI: Technology Architecture Is Now the “Risk Architecture” of a Bank
Reserve Bank of India Deputy Governor Rohit Jain delivered an important regulatory message at the SBI Banking & Economics Conclave in Mumbai on 24 September: technology risk should be treated as a first-order enterprise risk, not merely an IT issue.
Core banking, payments, APIs, cloud, data centres, fraud monitoring and regulatory reporting have become so interconnected that a technology failure can disrupt a financially healthy bank.
On AI, Jain said it can improve fraud detection, risk assessment and productivity, but can also amplify errors affecting credit, fraud alerts, pricing and customer access. His central message was particularly significant: “governance must precede scale.”
Boards and senior management need sufficient technology understanding to challenge assumptions, examine dependencies and test resilience.
Why it matters: This effectively pushes Cyber Risk + AI Risk + Third-Party Risk + Operational Resilience into board-level BFSI governance. For CISOs, CROs and compliance officers, technology investment increasingly becomes risk investment; AI models should have ownership, validation, monitoring, human override and auditable decision trails.
2. UP STF Arrests Alleged Exam Hacker “Devil”; Proxy Servers Enabled Remote Control of Candidates’ Computers
Uttar Pradesh Special Task Force has arrested Arun Kumar, alias “Devil,” in connection with alleged manipulation of the SSC GD Constable online examination.
Investigators say examination-server source files were decoded and modified so candidates’ computers inside examination centres could be connected through proxy infrastructure and controlled by remote solvers outside. The allegations remain subject to investigation and trial.
Investigators seized 11 Raspberry Pi devices, two laptops, 12 open TFT screens, VGA switches and other technical equipment. UP STF says the investigation has widened into a network involving examination centres, technical operators and people across several states.
Why it matters: This is much more sophisticated than conventional impersonation or question-paper leakage. It represents cyber-enabled examination manipulation involving endpoint access, server/software modification and remote-control infrastructure.
DFIR should reconstruct server changes, network traffic, Raspberry Pi configurations, proxy logs, remote-access sessions, device timestamps and financial links between candidates, centres and operators.
3. ED Arrests Two More in ₹146-Crore HashPe Cryptocurrency Fraud Investigation
The Enforcement Directorate has arrested Syed Usman alias Babu and Dhamodharan Balachandran in its PMLA investigation into the alleged HashPe cryptocurrency investment fraud. The arrests were made on 22 September and confirmed by ED on 24 September; both were sent to three days’ judicial custody.
Investigators allege that investors were induced to put money into HashPe with promises of high returns involving a purported cryptocurrency called TCX. Reporting on the investigation says approximately ₹146 crore was collected through cash, bank transfers and TCX coins.
ED alleges Usman handled cash and bank transfers while Balachandran managed software and wallet functions. These remain allegations and have not been judicially established.
Why it matters: The case illustrates the convergence of Cyber Fraud + Crypto + PMLA + Blockchain Forensics. Investigators need to combine bank trails with wallet clustering, exchange/KYC records, device forensics, software logs, token issuance records and beneficial-ownership analysis.
The investigative model increasingly becomes:
Victim → Bank/Cash → Crypto Token → Wallet → Exchange/Layering → Beneficiary → Asset → PMLA Attachment
Proposal for Conducting Cyber Crisis Drill, Tabletop Exercise (TTEx) & CCMP Readiness Exercise
4. Indian Army Outlines AI-Enabled Battlefield: Soldiers Could Command Multiple Unmanned Systems
Senior Indian Army leadership has provided one of its clearest descriptions yet of how AI could transform future warfare.
Lt Gen Puneet Ahuja, Director General Strategic Planning, said AI would change how soldiers fight rather than replace them, with applications spanning target acquisition, surveillance, logistics, predictive maintenance, decision support and cyber defence.
Future soldiers may control multiple unmanned platforms while simultaneously operating in an environment involving drones, electronic warfare, GPS denial, cyber disruption and autonomous systems.
Separately, CDS General N.S. Raja Subramani stressed Compute, Compatibility and Collaboration and said military AI must continue working when battlefield communications are degraded.
Why it matters: The decisive issue is not simply acquiring AI software. Battlefield AI must be secure, resilient, explainable where necessary and capable of operating in contested or offline environments.
It also creates a new DFIR discipline: after an autonomous-system incident, investigators may need to reconstruct sensor inputs, model decisions, operator commands, communications and software state.
5. OpenAI Reportedly Preparing GPT-6 Cyber as Specialised AI-Cybersecurity Race Accelerates
OpenAI is reportedly preparing to preview a cybersecurity-focused model called GPT-6 Cyber within days, according to a Fortune report cited by Reuters. The report says a limited group of customers in an application-only cybersecurity programme already has alpha access.
OpenAI had not confirmed the report to Reuters at publication, so the launch details should be treated as reported plans rather than an announced product release.
The reported model would arrive as frontier AI systems become increasingly capable of vulnerability research, code analysis and autonomous security work—while recent incidents have simultaneously demonstrated the danger of agents crossing intended network boundaries.
Why it matters: Cybersecurity appears to be entering an AI-vs-AI phase. Specialised cyber models could accelerate SOC triage, malware analysis, vulnerability discovery, secure-code review and penetration testing—but the same capability creates dual-use risk.
For police, defence and enterprise SOCs, the emerging stack could become:
Human Analyst → Cyber AI Agent → Tool/API Access → Automated Investigation → Human Approval → Response
The critical controls will be sandboxing, target allow-lists, least privilege, immutable agent logs and human authorisation for consequential actions.
Today’s Strategic Signal
A striking theme connects all five stories: software is becoming operational power.
RBI says a bank’s technology architecture is becoming its risk architecture; examination fraudsters allegedly manipulated servers rather than merely candidates; crypto fraud combines software and financial infrastructure; soldiers are moving toward human-machine teaming; and specialised AI may soon automate substantial portions of cybersecurity work.
For policing, BFSI and national security, the emerging requirement is therefore:
Technology Governance + Cyber Resilience + DFIR + AI Assurance + Human Accountability
Follow for daily updates on cybercrime, corporate fraud, DFIR, hacking, investigations, and digital forensics
