New US Weapons Supplier Causes Uproar as Senior Executive Suddenly Departs
Jakarta, CNBC Indonesia – The United States and Israel’s conflict with Iran has sparked another dispute, this time concerning an American defence contractor partner for AI tools used in military operations. Anthropic, which supplies AI tools to the US Department of Defence (DoD), has refused to allow its AI technology to be used for creating autonomous weapons or surveilling American citizens.
This stance infuriated the Trump administration, which immediately prohibited federal agencies from using Anthropic’s AI tools. Nevertheless, according to reports from the Wall Street Journal, Claude AI from Anthropic is still being used in operations against Iran for intelligence research and targeting adversaries.
Trump’s ban on Anthropic usage opened the door for competing service provider OpenAI. The ChatGPT creator announced it has newly signed a contract with the DoD to become the supplier of new AI tools for military purposes.
OpenAI’s decision immediately triggered controversy. Users began deleting the ChatGPT application from their mobile phones en masse and reportedly switched to Claude AI.
Beyond losing its user base, OpenAI also lost its leader. Caitlin Kalinowski, who headed hardware development at OpenAI, announced her resignation on Saturday, 7 March local time.
She stated that her reason for resignation was concern regarding OpenAI’s agreement with the DoD, according to Reuters reporting on Monday, 9 March 2026.
In a post on social media platform X, Kalinowski wrote that OpenAI rushed into approving the use of its AI models for the Pentagon’s classified cloud network.
“AI has an important role in national security. However, surveillance of American citizens without judicial oversight and lethal autonomy without human authorisation are boundaries that deserve further consideration,” Kalinowski wrote.
Reuters could not immediately request comment from Kalinowski. However, she wrote on X that she deeply respects OpenAI CEO Sam Altman and the entire OpenAI team.
She merely regretted that OpenAI announced the Pentagon agreement without clear restrictions.
“This is a governance issue. These matters are too important for rushed agreements or announcements,” she said.
OpenAI stated one day after the agreement was reached that the agreement includes additional safeguards to protect its use cases.
The company reaffirmed over the past weekend that their “red lines” prohibit the use of technology in domestic surveillance or autonomous weapons.
“We recognise that people have strong views on this matter and we will continue to engage in discussions with employees, government, civil society, and communities around the world,” the company said in a statement to Reuters.
Kalinowski joined OpenAI in 2024 after leading the development of augmented reality hardware at Meta Platforms.