# Helsing launches Clarity AI platform for target selection

> The German defense firm unveiled an agent‑based system that compiles sensor data into target recommendations, with every strike still requiring a human sign‑off.

Oossa · 2026-11-11 · https://oossa.com/en/helsing-launches-clarity-ai-platform-for-target-selection

On 11 November 2026 Helsing introduced Clarity at a press event in Berlin. The software bundles hundreds of specialized AI agents that pull together sensor feeds, open‑source intel, written reports and existing military systems. The agents match objects to a shared military ontology, rank them, and hand over a recommendation to a human operator. Only after a person authorizes the handoff does the system forward the target to a weapon platform. Helsing says each recommendation is logged with the name of the approving officer.

Clarity runs on the customer’s own hardware and uses models built in Europe, according to the company. It follows NATO’s Modular Open Systems Approach, so it can plug into weapons from any manufacturer. Helsing stresses that data and algorithms stay under the armed forces’ sole control, a point it highlights as a safeguard against foreign dependence.

## How it fits into the wider market

The platform enters a field dominated by U.S. firms such as Palantir, whose Maven system is already embedded in Pentagon and NATO workflows. Helsing is one of the few European vendors that can point to combat experience in Ukraine, where it supplies drones, though it did not say whether any customers are already using Clarity. Germany’s Bundeswehr has announced plans to field AI‑powered tools starting in 2027, which could open a market for systems like Clarity.

## Risks and open questions

Recent incidents show that AI‑driven targeting can go wrong. In Iran, U.S. forces used Palantir‑Claude pipelines that suggested thousands of targets in hours, prompting worries about “the computer said to do this.” In Gaza, a separate system flagged tens of thousands of civilians, with human checks lasting only about 20 seconds per target. A 2026 near‑miss involving a U.S. chatbot falsely reported a Chinese ship carried nuclear components illustrates how fused data can still produce hallucinations. Helsing’s promise of human‑named accountability aims to address these concerns, but the company did not detail verification procedures or independent testing of Clarity.

## The facts

- Helsing unveiled the Clarity platform in Berlin on 11 November 2026.
- Clarity processes sensor feeds, open‑source intel, written reports and existing military system data.
- The system uses hundreds of AI agents trained on real, synthetic and simulated data.
- Helsing says every target recommendation must be approved by a named human before weapon release.
- Germany’s Bundeswehr plans to start introducing AI‑enabled systems in 2027.

## Why it matters

For European militaries, Clarity offers a home‑grown alternative to U.S. targeting software, promising data sovereignty and traceable human responsibility. However, without clear standards for verifying AI output, the risk of accidental or unlawful strikes remains, and commanders will need robust review processes to avoid past mistakes.

## Sources & references

1. [Helsing built an AI system that prepares targets for weapons systems and says every decision will carry a human name](https://the-decoder.com/helsing-built-an-ai-system-that-prepares-targets-for-weapons-systems-and-says-every-decision-will-carry-a-human-name/) – The Decoder, 2026-11-11

Last updated: 2026-11-11