qwant news | Palantir’s ’Workflow’ of AI-Directed Death
Palantir’s “Maven Smart System,” showcased in a Pentagon video, is an AI‑driven target‑and‑destroy platform that lets operators move a detection through a single visualization tool with a few mouse clicks. Once a target is identified, the system automatically advances it to a “course of action” (CoA) workflow, where AI suggests the optimal weapon and the operator can immediately execute the strike—all within the same interface. The Pentagon touts this as a revolutionary consolidation of what previously required eight or nine separate systems, promising faster and more efficient kill‑chain closure.
The company insists a human remains “in the loop,” making the final decision to fire, a claim echoed by Palantir’s UK head to the BBC. The rollout comes after Google abandoned its own Project Maven eight years earlier amid employee protests over creating a “Google‑Earth for war.” More recently, Anthropic publicly resisted Pentagon requests to weaponize its Claude AI, only to be slapped with a federal‑contract ban by former President Donald Trump, who framed the dissent as a betrayal. These tensions highlight growing unease about autonomous weapons and the potential for AI systems to operate with minimal human oversight.
In the wake of the United States‑Israel strike that unintentionally devastated an Iranian elementary school, journalists allege Maven was likely involved in selecting that target, though official denial remains absent. The system’s integration into NATO and its suspected role in the thousands of missile strikes during the ongoing Iran conflict raise serious questions about accountability and the moral ramifications of delegating lethal decisions to AI. Truthdig’s article concludes by reaffirming its commitment to independent, fact‑based journalism amid the turmoil.
Read more: https://www.truthdig.com/articles/palantirs-workflow-of-ai-directed-death/





