OSINT Pulse February 2026 I Archives Under Scrutiny, New Tools, and What Voice AI Changes

Hyderabad: February brought a mix of disruption, new utilities and research insights for the OSINT community.

Wikipedia’s decision to blacklist Archive.today has raised fresh questions about the reliability of archiving platforms. At the same time, platforms like ShadowDragon continue to expand access to lightweight investigative tools, while Henk van Ess has released a practical crowd estimation utility. Adding to this, new academic research explores how voice-based AI interactions influence user behaviour.

Here’s a closer look at the key developments shaping investigative workflows.

Wikipedia bans Archive.today over security and integrity concerns

Wikipedia has blacklisted Archive.today following concerns raised within its editor community about safety and reliability. The move affects hundreds of thousands of links that previously pointed to the service as a backup source.

The decision followed reports that Archive.today embedded code within its CAPTCHA page that allegedly caused visitors’ browsers to send repeated requests to a security researcher’s website. The issue was that users may not have been aware that their browsers were generating that traffic. Editors also pointed to instances where archived pages appeared to have been modified after capture. That raised questions about whether the platform could be treated as a fixed and verifiable record of what was originally published.

For journalists and OSINT practitioners, archived links are frequently used to document deleted statements, altered webpages or content that later goes offline. When an archiving platform faces questions about how it handles user traffic or whether its snapshots remain unchanged, it affects how those links can be cited and relied upon in investigations.

At the same time, web content continues to disappear.

Articles are updated without clear revision histories, public pages are removed and entire domains shut down. Archiving remains a routine part of verification workflows, but this episode shows that the credibility of the archive itself can also become part of the scrutiny.

ShadowDragon and its free OSINT tools

ShadowDragon is a web-based OSINT platform designed for investigators, analysts and security teams to collect, correlate and analyse publicly available data.

Alongside its commercial products, ShadowDragon maintains a section on its website dedicated to free OSINT tools. These are lightweight utilities intended for everyday investigative tasks and are accessible without purchasing the full platform.

Some of the free tools include:

Email Permutator – Generates possible email address variations based on a person’s name or username, useful for account discovery and breach searches.

Dork Assistant – Helps build structured advanced search queries for search engines, supporting more precise discovery through operators.

OSINT Checklist Generator – Creates customisable investigation checklists to structure research workflows.

Image Forensics Tool – Provides basic image analysis features to detect potential manipulation.

Open Sources Toolkit – A curated collection of publicly accessible investigative resources.

While the enterprise platform focuses on large-scale data integration and link analysis, the free tools serve as practical utilities for independent researchers and journalists who need quick investigative support without complex infrastructure.

CrowdCounter by Henk van Ess

Henk van Ess has introduced CrowdCounter, a free desktop-based tool designed to estimate the number of people in a crowd photograph. The tool applies the Jacobs Method, a widely used crowd estimation technique that calculates attendance based on area size and density.

Users begin by defining the total area shown in the image. After uploading a photo, they select grid cells where people are visible and indicate how densely packed the crowd appears. Based on these inputs, the tool generates an estimated minimum and maximum range to reflect variation in crowd density.

CrowdCounter does not require installation or user registration and is intended to assist with basic visual crowd size estimation from images. However, the method appears more suitable for enclosed or clearly bounded spaces where area measurement is easier to approximate. Estimating crowds in large open grounds, uneven terrain, or political rally settings where density fluctuates significantly may present additional limitations.

Voice, perception and politeness in AI conversations

A recent study published in the International Journal of Human–Computer Interaction examines whether people behave more politely toward AI when they perceive it as more ‘conscious.’ Titled Do You Say “Please” to ChatGPT?, the research explores how interaction mode influences both perceived AI consciousness and user politeness.

The researchers conducted controlled experiments with 25 participants who interacted with ChatGPT in both text and voice modes. They measured how conscious users perceived the AI to be and how politely users behaved during conversations.

The findings show that participants rated ChatGPT as more conscious and more human-like in voice interactions compared to text. Voice mode increased the sense of awareness and social presence.

Users were also more polite in voice interactions. They used more courteous expressions, avoided harsh commands and maintained a more respectful tone. However, this increase in politeness was not linked to how conscious they believed the AI to be. Statistical analysis found no significant relationship between perceived AI consciousness and politeness levels. Instead, the interaction mode itself influenced behaviour, with voice conversations encouraging more socially appropriate language than text-based exchanges.

The study did not examine whether politeness changed the quality or content of the AI’s responses. The focus remained on user perception and behaviour rather than differences in output.

That’s all for February’s OSINT Pulse.

See you next month!

Dheeshma Puzhakkal

(OSINT Pulse is a monthly report by Dheeshma Puzhakkal, Editor of NewsMeter’s Fact Check team. The column tracks emerging developments in OSINT and AI, with a focus on what matters to Indian readers and OSINT professionals. For comments, insights, or leads, write to dheeshma.p@newsmeter.in. NewsMeter has no financial relationship with any of the companies or tools mentioned in this series.)