OpenAI Hardware Chief Resigns Over Pentagon Deal

The ethical fault lines within the artificial intelligence sector have shifted dramatically following the high-profile resignation of Caitlin Kalinowski, OpenAI’s Head of Hardware. Her departure comes as a direct protest against a controversial contract signed between the Sam Altman-led organisation and the Pentagon, the United States Department of Defense. The deal, which integrates OpenAI’s proprietary models into classified government networks, has ignited a firestorm of criticism both internally and globally.

A Principled Exit

Kalinowski, a heavyweight in the tech industry who previously spearheaded augmented reality (AR) glasses at Meta, joined OpenAI in November 2024 to lead their ambitious robotics and hardware ventures. However, her tenure was cut short by what she describes as a fundamental misalignment of values regarding military surveillance.

In her resignation statement, Kalinowski acknowledged that AI undeniably possesses a critical role in national security. However, she raised grave concerns regarding “surveillance of Americans without judicial oversight” and the terrifying prospect of AI-driven decision-making “without human intervention.” Her exit highlights a growing rift among Silicon Valley elite who fear that rapid militarisation of AI could bypass essential legal and ethical safeguards.

Public Backlash and Commercial Fallout

The revelation of the Pentagon partnership has not only caused internal strife but has also hit OpenAI’s bottom line. Since the contract was made public last week, social media campaigns advocating for the boycott of OpenAI products have gained significant traction. Preliminary reports suggest that hundreds of thousands of users have already cancelled their ChatGPT Plus subscriptions in a show of solidarity with those opposing military AI applications.

Key Milestones in OpenAI’s Defence Shift

The “Dual-Use” Dilemma

While OpenAI maintains that its technology will be used for administrative, logistical, and cybersecurity tasks within the Pentagon, critics argue that the distinction between “support” and “active warfare” is dangerously thin. By providing the digital infrastructure for the US military, OpenAI is seen by many as becoming a de facto defence contractor, a move that contradicts its original founding mission to ensure AGI benefits “all of humanity.”

As the hardware division loses its visionary leader, the future of OpenAI’s physical tech—including its rumoured robotic assistants—remains shrouded in uncertainty.