The goal of artificial intelligence is to minimize human error.
Palantir Technologies(PLTR) Additional artificial intelligence was added in response to the tragic U.S. target failure in one of the most high-profile attacks in the Iran conflict.
CBS News and Bloomberg reported that a U.S. missile attack on Shajala Tayebe Elementary School in Minab on February 28 killed 123 children. Pentagon investigators blamed the attack on intelligence flaws, outdated imagery and overreliance on artificial intelligence, Bloomberg reported.
Palantir chief technology officer Shyam Sankar disputed some of the claims.
Speaking on CBS’s “Face the Nation” on Oct. 11, Sankar said “the likely cause is human error” and warned that the Pentagon’s final report had not yet been released. He also revealed what Palantir has done since: It built additional AI agents to independently check more targeting efforts.
Because of these factors, the event was far more significant than discussing a single military action.
Palantir is betting that remedies for mistakes made by AI-assisted systems may increasingly involve another AI system observing the first one.
Palantir says the problem isn’t its artificial intelligence
One of the most important options that a computer system can assist with is Project Maven.
By integrating data from disparate military and intelligence sources, the platform helps staff identify targets, organize operations and advance operations through the military approval process. According to Palantir, the Maven Intelligence System is an artificial intelligence command and control platform used by combatant commands and U.S. military departments.
This conclusion does not mean that Maven selects the content it bombs on its own.
Also read: Palantir hits new record, its next AI market becomes awkward
Sankar stressed that humans remain responsible for fatal decisions. He told CBS that military personnel could use the system to check whether a potential target is on a no-hit list, is close to civilian infrastructure and whether other information raises concerns.
The Minab attack tested this framework.
The site was previously identified as an Iranian military installation, but images show that part of the site has long been used as a school, according to Bloomberg. Investigators determined that outdated information still existed in the intelligence chain, and some personnel wanted Maven to identify outdated or inconsistent data, the report said.
Palantir told Bloomberg that its software did not trigger the attack and that it was not responsible for the underlying intelligence.
Sankar continued on Sunday.
He described Bloomberg’s assertion that staff were too reliant on Maven as “inaccurate” and said the outcome of the Pentagon’s final review should be known. According to CBS News, the difference is important.
Palantir doesn’t want government customers or investors to think the error was caused by its software.
However, the company’s subsequent actions may be more revealing.
Palantir builds artificial intelligence agents to inspect target processes
Sankar said Palantir quickly created additional AI agents to perform more checks outside of the parts of the target process that Palantir directly controlled.
In other words, Palantir provides an extra pair of digital eyes.
New technologies may re-examine the data before an attack is carried out to identify discrepancies or elements that disqualify the target. According to Bloomberg, Palantir created skills to reassess intelligence and identify issues that human reviewers might have missed. According to CBS News, this approach has led to startling contradictions.
Palantir’s solution is to use artificial intelligence to arrest people if human error leads to a disaster.
If overreliance on AI is a problem, the solution still involves additional AI, but this time, one system assists in auditing another system.
More Palantir:
- Palantir’s latest AI initiative reveals bigger ambitions
- Palantir faces $330 million test, U.S. investors should pay close attention
- UBS resets Palantir stock price target for remainder of 2026
This is how artificial intelligence is increasingly finding its way into sensitive industries.
A bank might use one AI system to approve transactions and another to look for fraud.
A hospital might use software to suggest a diagnosis, while another system checks for contraindications.
On the battlefield, the stakes become extremely high.
The question is no longer just whether artificial intelligence should be involved in decision-making.
It became One question is how many layers of AI should check each other before humans act.
The Pentagon is digging deeper into the Palantir system
The controversy comes as Maven becomes more, not less, important to Palantir’s government business.
The Pentagon moved earlier this year to establish the Maven Intelligence System as an official program of record, giving the platform a more permanent place in U.S. military planning and funding. The system is already in use across U.S. combatant commands.
That makes the Minab incident particularly important for investors.
In the second quarter, Palantir’s U.S. government revenue increased 90% year-on-year to US$809 million, and total revenue increased 93% to US$1.94 billion.
These numbers illustrate why companies don’t view battlefield AI as an experimental side project.
Palantir claims software can reduce complex judgments from hours or days to minutes, with one of the best examples being Maven, although government work remains an important pillar of the company’s operations.
In combat, this speed is useful.
This is also where the danger occurs.
Military commanders may benefit from faster targeting, but it also reduces the time it takes for people to identify outdated information, conflicting facts or assumptions hidden in software.
The Minab investigation is therefore a test of the fundamental promise of Maven and military artificial intelligence: that automation can increase speed without compromising judgment.
Bloomberg/Getty Images
Palantir Now Faces Artificial Intelligence’s Toughest Problem
Beyond defense, this narrative is important for another reason.
Most commercial AI failures are reversible.
Chatbots provide poor responses. A business discovers an inaccurate spreadsheet. Recommendation engines can recommend incorrect products.
The goals of the military are different.
The software cannot undo the selection made.
The disagreement over the Minab incident is therefore crucial. The Bloomberg investigation highlighted flaws such as outdated intelligence, human processes and overreliance on artificial intelligence.
Sankar claimed that the information about Maven was incorrect and suggested that human error was likely to be the culprit. The Pentagon has not yet released the final results of the study.
Congress has asked questions.
In a March letter to Defense Secretary Pete Hegseth, more than 120 House Democrats requested details about the school strikes, the targeting process and Maven’s involvement. They also demanded that the findings of the investigation be made public.
Sanka told CBS that Palantir has briefed Congress on Maven’s capabilities but deferred questions about the Minab investigation itself to the Defense Department.
This leaves a key question unanswered.
When something goes wrong with an AI-assisted process, who is responsible for the mistake?
Software company?
The one who operates it?
Provide intelligence to the system?
Or the commander who ultimately approved the decision?
Palantir’s business will increasingly depend on whether governments believe those areas of responsibility remain clear.
What Palantir investors should watch next
Whether Palantir is responsible for the Minab strike may not be the most important development.
The buildings the company built in response to the strike were the most important developments.
Palantir added that artificial intelligence was designed to review more data from the targeting process for the incident. This means that autonomous agents performing tasks may not be the only aspect of the next phase of artificial intelligence.
Agents spying on each other, software verifiers, and people choosing which to trust may all be part of this.
This could lead to another business opportunity for Palantir.
Its software already sits between decision makers and vast amounts of data. If consumers start needing a second layer of AI to verify whether the first layer missed something, Palantir can sell not only faster decision-making but also the controls around it.
This has the potential to be very effective.
It’s also a greater responsibility.
For years, Palantir has argued that artificial intelligence should leave demos and enter the real world.
The real world demonstrates exactly the risks Minab faces.
Palantir isn’t backing away from artificial intelligence.
The aim is to build another one.
Related: Goldman Sachs resets 12-month price target on Palantir stock by 18%