Artificial intelligence could fundamentally change intelligence work — not because states suddenly collect more information, but because they may finally be able to understand almost everything they already collect.
In Episode 4 of Strategic Conversations with Helge Luras, former Norwegian intelligence analyst Helge Luras explores how AI and large language models may transform surveillance, intelligence collection and analysis.
Historically, intelligence agencies could intercept enormous amounts of communications, documents and other data, but human attention was the bottleneck. Analysts simply could not listen to every conversation, read every message or connect every relevant data point.
AI changes that.
A system can potentially transcribe, translate, classify and search massive volumes of information, reconstruct networks of relationships and identify patterns that would previously have taken thousands of human hours to uncover. Information collected years earlier may also gain new value if AI can retrospectively connect it to a person or event that only later becomes significant.
But greater analytical power creates new vulnerabilities.
If intelligence services increasingly rely on AI to detect patterns, adversaries may respond by manufacturing false ones — synthetic identities, deepfakes, fake communications and deliberate disinformation designed to contaminate the opponent’s picture of reality.
The future of intelligence could therefore become a contest not only over who collects the most information, but over who can still determine what is real.
The conversation then turns to domestic surveillance.
Authoritarian states may have fewer legal or political constraints on collecting communications, financial data, camera footage and other information about their own populations. AI could make such systems extraordinarily powerful.
Could that give authoritarian states an economic or strategic advantage over more open societies?
Or does pervasive surveillance create another problem: people who believe they are constantly monitored may become more cautious, conformist and unwilling to speak honestly, experiment or criticize authority?
Luras then raises a deeper concern: even democratic societies may find these capabilities increasingly difficult to resist.
Security agencies can always argue that more data may be needed to stop terrorism, espionage or foreign interference. Once data are collected, and AI dramatically lowers the cost of analyzing them, how realistic is it to expect governments never to use that capability?
The discussion ends with a warning about a possible ratchet effect: each new threat may justify greater surveillance powers, while the technical and institutional infrastructure created in response becomes difficult to dismantle later.
Ultimately, the episode asks:
Can liberal democracies preserve meaningful privacy in a world where technology increasingly makes total surveillance possible?
Strategic Conversations with Helge Luras uses artificial intelligence as an intellectual sparring partner to explore geopolitics, technology, power and the forces reshaping modern society.
Host: Helge Luras
AI conversation partner: ChatGPT