AI Intent Indicator Autor: Libor Benes

AI Intent Indicator is a Firefox browser sidebar extension that performs deterministic, rule-based analysis of linguistic intent indicators in AI-related text.

Critical Distinction: This tool highlights linguistic indicators, it does NOT detect, infer, or classify intent. The description "Indicator" (not "Detector") is deliberate and precise.

Purpose:
The tool detects and highlights linguistic patterns that may indicate different types of communicative intent in AI-related text:
• Instruction: Direct commands and procedures.
• Obligation: Mandatory requirements and duties.
• Suggestion: Recommendations and options.
• Uncertainty: Hesitation and possibilities.
• Warning: Cautions and risks.
• Prohibition: Things to avoid and restrictions.

What This Tool Does:
• Accepts any text input for analysis.
• Detects linguistic INDICATORS using explicit, transparent pattern matching.
• Highlights detected indicators with color coding across 6 categories.
• Counts occurrences per category with colored summary badges.
• Produces reproducible, explainable results.
• Works fully offline with no external dependencies.
• Uses deterministic pattern matching (NO machine learning or AI inference).
• Supports export of highlighted text for documentation.

What This Tool Does NOT Do:
• Does NOT detect, infer, or classify intent.
• Does NOT interpret or understand meaning.
• Does NOT judge text quality or effectiveness.
• Does NOT predict AI behavior or responses.
• Does NOT infer implicit meaning or context.
• Does NOT modify text automatically.
• Does NOT use AI or machine learning inference.

Limitations & Caveats:
• Indicators, not intent detection: This tool highlights linguistic INDICATORS to assist human review. It does NOT detect, infer, or classify intent. Results are indicative, not exhaustive or definitive.
• Ambiguity inherent in language: Ambiguity is a property of natural language. Some words may legitimately indicate multiple categories, and some indicators may be missed or misclassified.
• Static, rule-based analysis: The analysis is static and rule-based. It evaluates visible wording only and does not infer context, tone, compliance, or speaker motivation.
• Word form coverage: Rules account for common grammatical variations (e.g., tense, plurality, capitalization). Not all linguistic forms or synonyms are guaranteed to be detected.
• No guarantee of accuracy or completeness: This is a language analysis aid. Even dynamic language analysis tools misfire occasionally. Use as a starting point for review, not as a definitive analysis.
• Heuristic signals only: Indicators are heuristic signals only. Apply your own judgment and critical thinking. The tool provides data points, not conclusions.

Pattern Coverage:
The tool recognizes extensive linguistic patterns in U.S. English and captures common word forms:
• Instruction Patterns (150+ patterns with forms): Direct commands, verification procedures, creation instructions, analysis directives.
• Obligation Patterns (100+ patterns with forms): Mandatory requirements, duties, expectations, commitments.
• Suggestion Patterns (80+ patterns with forms): Recommendations, options, considerations, potential actions.
• Uncertainty Patterns (70+ patterns with forms): Hesitation, ambiguity, possibilities, likelihoods, speculation.
• Warning Patterns (80+ patterns with forms): Cautions, dangers, risks, problems, critical issues.
• Prohibition Patterns (70+ patterns with forms): Restrictions, bans, avoidance, prevention, limitations.

Security & Privacy:
🔒 Security-First Design:
Unlike web tools that require internet, might track your text or use unreliable AI inference, AI Intent Indicator is:
• Private: 100% offline, no data collection, no telemetry.
• Deterministic: Same input → same output every time.
• Professional: Designed for critical review, not casual use.
• Transparent: Detection patterns are indicated in category tooltips. If a word isn't highlighted, it's not in the rule set. No hidden logic or unknowns. The tool makes no claims of completeness.

• Minimal permissions.
• No network connections: Everything bundled locally, 100% offline.
• No tracking or telemetry: Zero analytics, 100% private.
• No data collection: Your text never leaves your browser.
• Secure implementation and coding.
• Local processing: All analysis happens in your browser.
• Privacy by design: Built for air-gapped, high-security environments.

Target Audience:
AI Intent Indicator is designed for professional users who need to analyze linguistic patterns in:
• AI prompt engineering: Reviewing prompt structure and directive language.
• AI governance: Analyzing policies, guidelines, and requirements.
• Content review: Assessing AI-generated text for directive patterns.
• Security analysis: Identifying command language in AI interactions.
• Research: Studying linguistic patterns in AI communication.