Uncategorized

“IRIS” (Intelligent-Rapid Investigation System): The AI-Augmented OSINT Framework

In Taqasee, We’re constantly exploring new ways to enhance our OSINT (Open-Source Intelligence) capabilities. The emergence of powerful AI tools presents a significant opportunity to revolutionize our workflows. Here’s a framework we’ve conceptualized that integrates Perplexity AI, NotebookLM, and AI agents like Genspark to augment our OSINT investigations.

This framework is designed to be iterative and flexible, allowing investigators to move between stages as new information emerges.

IRIS Framework Stages

Phase 1: Discovery & Broad Understanding (Perplexity AI & AI Agents – Genspark)

Objective: Quickly grasp the scope of an event, identify key actors, and uncover initial leads.
How to use it:
- Perplexity AI: Start by asking broad, natural language questions about the incident or topic. Perplexity’s strength lies in its ability to synthesize information from various web sources and provide summarized answers with citations. This helps in getting a quick overview and identifying initial keywords, entities, and potential data points.
- AI Agents (e.g., Genspark): Deploy AI agents to conduct wider, more automated searches based on the initial findings from Perplexity. Genspark, for instance, can create “Sparkpages” – custom-generated knowledge pages – by synthesizing information from multiple trusted sources, potentially uncovering less obvious connections or pieces of information. These agents can be tasked to monitor specific keywords or entities across various platforms.

Phase 2: Deep Dive & Focused Research (Perplexity AI & Traditional OSINT Tools)

Objective: Drill down into specific leads, verify information, and gather detailed evidence.
How to use it:
- Perplexity AI: Use Perplexity for more targeted queries. Ask specific questions about individuals, organizations, locations, or technical details identified in Phase 1. Its ability to understand context and provide sourced answers makes it valuable for clarifying specific points.

Phase 3: Data Organization & Analysis (NotebookLM)

Objective: Consolidate all gathered information, identify patterns, connect disparate data points, and build a coherent narrative.
How to use it:
- NotebookLM: This is your central AI-powered digital notebook. Upload all relevant documents, articles, social media captures, notes, and multimedia files gathered in the previous phases.
- Summarization & Q&A: Use NotebookLM’s AI capabilities to summarize large documents, get quick answers from your uploaded sources, and identify key entities and relationships within your data.Unlike generative AIs that might invent answers, NotebookLM grounds its output in the user-provided data, enhancing reliability.
- Connections & Insights: Leverage NotebookLM to help map relationships and interactions between individuals and organizations. Its AI can help you see connections you might have missed and generate hypotheses based only on the information you’ve fed it.
- AI-Generated Podcasts/Audio Overviews: A newer feature of NotebookLM, “Audio Overviews,” can turn your documents and sources into a conversational discussion between AI hosts. This can be a novel way to review and digest complex information or even to brief team members.

Phase 4: Hypothesis Generation & Testing (AI Agents – Genspark & Perplexity AI)

Objective: Formulate and test investigative hypotheses based on the analyzed data.
How to use it:
- AI Agents (e.g., Genspark): Based on the patterns and connections identified in NotebookLM, use AI agents to proactively search for new information that could confirm or refute your hypothesis. For example, if you hypothesize a specific individual was at a certain location, an AI agent could be tasked to continuously monitor for any new publicly available data related to that individual and location.
- Perplexity AI: Use Perplexity to explore “what-if” scenarios or to find information that supports or contradicts your working theories. Frame your questions to challenge your assumptions.

Phase 5: Reporting & Dissemination (NotebookLM & Human Oversight)

Objective: Compile findings into a structured, evidence-based report.
How to use it:
NotebookLM: Utilize NotebookLM to help structure your report. Its summarization and data organization features can assist in drafting sections of your investigation. You can ask it to create outlines or generate summaries of key findings based on your uploaded sources.

Benefits of the IRIS Framework

Speed & Efficiency: AI tools can rapidly process vast amounts of data, allowing investigators to quickly identify relevant information and leads.
Enhanced Discovery: AI can uncover connections and patterns that might be missed by human researchers.
Improved Organization: Tools like NotebookLM provide a structured way to manage and analyze large, complex datasets.
Scalability: AI agents can automate repetitive search and monitoring tasks, freeing up human investigators for more complex analysis.

Ethical Considerations & Limitations

Bias in AI: AI models can reflect biases present in their training data. It’s crucial to be aware of this and critically evaluate AI-generated outputs.
Misinformation & Disinformation: AI can also be used to create sophisticated disinformation. Investigators must be vigilant in verifying the authenticity of sources.
Data Privacy: Ensure all data collection and analysis comply with legal and ethical guidelines. Some AI tools might involve sharing your data with third parties, so understanding their terms of service is essential.
Over-Reliance: AI tools should augment, not replace, human investigative skills, critical thinking, and final judgment.

This IRIS framework provides a structured yet adaptable approach to leveraging the power of Perplexity AI, NotebookLM, and AI agents like Genspark to enhance the speed, depth, and effectiveness of our OSINT investigations, while always upholding Taqasee’s core principles of accuracy and transparency.