AI Targeting May Outrun Human Control Before Disaster Strikes

Human approval may not provide meaningful control over military decisions influenced by artificial intelligence when automated systems move faster than people can evaluate their recommendations, witnesses told Congress Wednesday.

The warning extended beyond weapons capable of selecting and engaging targets without direct human action.

Witnesses said AI can shape the information and recommendations behind a strike while leaving a person formally responsible for approving the final decision.

“The approval button alone does not demonstrate the control,” Anna Mysyshyn, a Ukrainian AI governance researcher, told lawmakers during the hearing, titled “Artificial Intelligence in the Military Domain: Implications for Human Rights.”

Gold. Upgraded. Buy real gold from $1. Store it in America. Sell it with one click.

The Tom Lantos Human Rights Commission examined the issue as the Pentagon seeks to accelerate military decisions through AI.

The Pentagon’s 2023 AI adoption strategy identified “fast, precise and resilient kill chains” as a desired result.

Its 2026 AI strategy calls for the military to become an “AI-first” warfighting force by redesigning workflows and operational concepts around existing capabilities.

That push has sharpened questions about whether human supervision can remain effective as decision windows shrink.

This Could Be the Most Important Video Gun Owners Watch All Year

Image Credit: Lockheed Martin Skunk Works

Mysyshyn questioned whether one person supervising several AI assisted systems could detect an incorrect recommendation and intervene before harm occurs. She cited Ukraine as an example where AI has reduced decision timelines from hours to minutes.

She said a person making the final call may lack the information, time or ability needed to challenge a system’s recommendation.

The War Department’s 2023 directive on autonomy in weapon systems requires autonomous and semi autonomous weapons to preserve appropriate levels of human judgment over force.

Amanda Klasing, national director of government relations and advocacy at Amnesty International USA, warned that AI could compress the military “kill chain” until determining whether meaningful human discernment occurred before a strike becomes difficult.

Joseph Chapa, a former military officer and scholar of military ethics, said policymakers should look beyond a system’s degree of autonomy.

He urged them to examine statistical inference, failures within particular operational environments and whether decisions can be audited afterward.

Chapa said statistical AI can react to relatively small input changes with substantially different results.

Image Credit: DoW

Complex systems can also make reconstructing the process behind a recommendation difficult after an operation, which led him to call for broader AI education and specialized training for targeting and intelligence personnel.

Witnesses also raised the accountability problem created when an AI assisted decision contributes to civilian harm.

The War Department established its Civilian Harm Mitigation and Response Action Plan in 2022, including a Civilian Protection Center of Excellence that federal law requires the Pentagon to operate.

A War Department inspector general review released in May found personnel losses and reassignments, the end of funding for a civilian harm data management platform and the failure of a senior steering committee to continue meeting.

Klasing said the military currently has limited ability to determine whether AI contributes to improper civilian targeting.

On the first day of the war between the U.S. and Iran, a Tomahawk cruise missile struck Shajareh Tayyebeh elementary school in Minab, southern Iran.

At least 168 people were killed, including more than 100 children under age 12, according to UN and Iranian officials.

The school was fewer than 100 yards from an Islamic Revolutionary Guard Corps naval installation and had previously been inside the IRGC compound perimeter, according to Amnesty International analysis of satellite imagery.

A wall appeared between 2013 and 2016, and Reuters found the school was established years before the U.S. and Israel launched their first strikes on Feb. 28.

More than 120 House Democrats wrote to the Pentagon on March 12 seeking clarity about whether AI, including the Maven Smart System, helped identify the school as a target.

U.S. military investigators said the strike was likely carried out by U.S. forces, but the Pentagon has not released its investigative findings.

Klasing proposed including information about AI involvement in the Pentagon’s annual civilian casualty reports required under Section 1057 of the National Defense Authorization Act.

Existing requirements cover mitigation efforts and investigations but do not specifically require disclosure of whether AI contributed to an incident.

Steve Feldstein, a senior fellow at the Carnegie Endowment for International Peace, called for operational data collection and reviews examining target identification and civilian casualties.

“A bigger question is to what extent is the Pentagon collecting data and verifying the target accuracy of all these 1000s of strikes?” Feldstein said.

Feldstein also said AI can verify intelligence, detect imagery changes and prioritize defensive responses to drones and ballistic missiles.

But witnesses warned that AI cannot repair flawed human rules, including assumptions about lawful targets or acceptable civilian harm.

The commission cited a 2004 incident in Kirkuk, Iraq, where Army squad leader Nick Ellison saw two young men firing a machine gun from a car.

Ellison noticed panic, ordered his soldiers to hold fire and learned the men were brothers of a newly married groom firing in celebration.

Mysyshyn called for testing of bias, reliability, civilian protection, time pressure, operator intervention and decision tracing.

The central concern was whether a human approver has enough time, information and authority to make a meaningful choice before harm occurs.

Gold is the oldest answer. SoundMoney makes owning it as easy as the apps you already use.

The opinions expressed by contributors and/or content partners are their own and do not necessarily reflect the views of LifeZette. Contact us for guidelines on submitting your own commentary.