Caitlin Kalinowski, OpenAI’s robotics hardware lead, resigned in early March 2026 after the company agreed to deploy advanced AI systems in classified Pentagon environments. She said AI can have a role in national security, but objected to what she described as a rushed announcement without adequately defined safeguards for domestic surveillance and lethal autonomy. OpenAI says its agreement prohibits those uses; its public account does not independently establish how the safeguards will work in practice.
Who resigned, and what did she say?
Kalinowski led hardware work within OpenAI’s robotics organization. Engadget’s coverage was corrected to describe her as the company’s “robotics hardware lead,” rather than the broader “head of robotics” title used in some initial coverage. Engadget reported that she had previously worked at Meta and joined OpenAI in late 2024.
In her resignation statement, Kalinowski said she cared deeply about the robotics team and its work. She acknowledged that AI has an important role in national security, while arguing that domestic surveillance of Americans without judicial oversight and lethal autonomy without human authorization required more deliberation. She characterized her decision as “about principle, not people.” In a follow-up clarification, she said her concern was governance: the agreement was announced before guardrails had been defined sufficiently.
OpenAI confirmed her resignation to Engadget and said it understood that people had strong views about the issues. The company said it did not support the prohibited uses Kalinowski identified and described its agreement as a responsible way to provide national-security AI. On March 8, 2026, OpenAI told Engadget it had no plans to replace her; that statement describes the company’s position at that time, not necessarily its current staffing.
#1 Best Overall
- BUILD, CODE & DRIVE YOUR OWN ROBOT CAR: Turn coding, electronics and engineering into a working programmable robot car you can assemble, program and drive; ideal for weekend family projects, STEM classrooms, coding clubs, robotics lessons and maker challenges
- EXPLORE FPV, LINE TRACKING & OBSTACLE AVOIDANCE: Control the robot with the ELEGOO app or IR remote, view live FPV video through the onboard camera, follow black lines, avoid obstacles with the ultrasonic sensor and explore multiple interactive driving modes
- BEGINNER-FRIENDLY BUILD WITH GUIDED WIRING: Keyed XH2.54 connectors help reduce wiring mistakes, while the illustrated tutorial and example programs guide beginners step by step from chassis assembly and module connection to programming and the first successful run
- GO BEYOND ASSEMBLY WITH CREATIVE CODING: Program with Arduino IDE to explore movement, sensors and control logic, then modify example code to create custom routes, reactions and robotics experiments that develop coding, problem-solving and engineering skills
- COMPLETE RECHARGEABLE STEM ROBOTICS KIT: Includes an ELEGOO UNO R3 controller board, ESP32-WROVER-based camera and Wi-Fi module, line-tracking and ultrasonic sensors, motors, IR remote and a 2000 mAh rechargeable lithium-ion battery; recommended for ages 8+ with adult guidance for first-time builders
What was the Pentagon agreement?
OpenAI announced an agreement on February 28, 2026, covering deployment of advanced AI systems in classified environments. Its official announcement is titled “Our agreement with the Department of War”; contemporary coverage commonly calls the counterparty the Department of Defense or the Pentagon. The terminology differs, but the announcement concerns the Pentagon institution.
OpenAI describes three central restrictions in the agreement:
Rank #2
- 35+ Guided Electronics Projects: Progress from LEDs and buttons to RFID access, real-time clocks, motion and distance sensing, environmental monitoring, motor control and interactive displays for STEM learning, coding clubs and maker projects
- More I/O and Memory for Larger Builds: The MEGA 2560 R3 provides 54 digital I/O pins, including 15 PWM outputs, 16 analog inputs, 4 hardware serial ports and 256 KB flash for projects that combine more sensors, controls and displays
- 200+ Components for Prototyping: Includes LCD1602, RC522 RFID, RTC, DHT11, HC-SR501 PIR, ultrasonic and water-level sensors, GY-521, MAX7219, keypad, joystick, rotary encoder, relay, SG90 servo, stepper motor, DC motor, breadboard and more
- Learn, Modify and Create: Follow 35+ guided lessons with example code, then adjust sensor thresholds, timing, display text, motor behavior and control logic to turn structured exercises into access systems, monitors, alarms and interactive projects
- Organized for Repeatable Learning: Pre-soldered modules, a solderless breadboard, storage case and small-parts box reduce setup time and keep sensors, LEDs, ICs, wires and other components easy to find between projects
- Domestic surveillance: OpenAI says its tools may not be intentionally used for mass domestic surveillance.
- Autonomous weapons: The company says its systems may not independently direct autonomous weapons in circumstances where law, regulation, or Department policy requires human control.
- High-stakes automated decisions: OpenAI says its systems may not make decisions that require human approval, giving social-credit-style determinations as an example.
OpenAI also says the arrangement is cloud-only: it will not provide “guardrails off” models or deploy its models directly on edge devices. The company says it will retain control of its safety stack and keep cleared OpenAI engineers and safety or alignment researchers involved. Those are the company’s descriptions of contractual and technical controls, not independent verification that the controls are sufficient or will be effective in every operational setting.
How the surveillance language changed
On March 2, OpenAI updated its public explanation with additional language on domestic surveillance. The company said the agreement explicitly barred intentional domestic surveillance of U.S. persons or nationals, including through procurement or use of commercially acquired personal or identifiable information. It also said the Department had affirmed that its intelligence agencies, such as the NSA, would not use OpenAI’s services without a new agreement.
Recommended Free Tools
Rank #3
- 🎁Ideal Gift for Kids & Teens: Celebrate child’s growing skills and important milestones with this 5-in-1 Programmable robot set. Whether for birthdays, holidays, or achievements, it’s the perfect gift that encourages learning and hands-on fun—a gift that grows with them
- ✨STEM Educational Toys: The robot set for kids ages 8+ combines the fun of STEM learning. It encourages hands-on learning and early programming as they build, which can spark creativity and imagination and provide hours of screen-free play
- 📱Flexible Dual Control Modes: Control the Robotic kit with the intuitive app (Bluetooth) or remote. Enjoy fun features like basic programming, path, and precise movement, exploring endless interactive play
- 🔄 5-in-1 Buildable with Varying Difficulty: The Robot Kit with Progressive Difficulty! From simple robots to complex models, kids can build a robot, dinosaur, car, tank, and more. Adjustable head, arms, and tail allow for fun, playful poses. Perfect for kids 8-12 to develop skills step by step and ignite creativity
- 🛠️Clear & Detailed Build Instructions: This robot kit includes 488 pieces, with clear, colorful step-by-step instructions to make assembly easy. Kids can build their own robots independently or with family, enjoying quality time together and a confidence-boosting building experience
The timing matters: OpenAI’s public account indicates that this language was added during follow-up discussions and described in the March 2 update, after the February 28 announcement. The public material does not establish that all of the expanded language was public when the agreement was first announced.
What the safeguards do—and do not—establish
Contract restrictions are not the same as technical guarantees
OpenAI presents the agreement as a combination of contractual limits, cloud deployment, continued safety systems, and staff involvement. Each addresses a different part of the risk. Contract terms define permitted use; architecture may limit how a model can be connected to systems; safety controls and personnel are intended to shape or review operation. The available public sources do not independently assess enforcement, monitoring, or remedies if the terms are breached, particularly in classified environments.
Rank #4
- 🎁 Ideal Gift for Kids & Teens: This STEM solar robot kit celebrates child’s growing skills and important milestones. Whether for birthdays, holidays, it’s the perfect gift that grows with them and offers screen-free fun
- 📚 STEM Educational Toy: This solar educational toy brings science to life! The fun DIY building experience sparks children's curiosity in engineering and renewable energy, while nurturing their problem-solving skills
- ☀️ Powered by the Sun: Enjoy outdoor play with solar power or switch to a strong artificial light source indoors, such as a flashlight, ensuring uninterrupted play for children. This solar build bot toy encourages kids to have fun while exploring renewable energy
- ⚡ Upgraded Larger Solar Panel: Features a large sun-catching surface to harvest more sunlight and deliver stronger power output. Kids discover renewable energy principles through play - a fun educational toy for ages 8+
- 🤖 12-in-1 Buildable with Increasing Challenge: With 190 parts, kids can build 12 models like robots, cars, and more. From simple beginners to advanced builds, the varying difficulty levels allow it to grow with your child’s skills. Each robot sparks children’s creativity
Human control and autonomous weapons
OpenAI’s restriction is not a blanket statement that its technology cannot be used in any military system. Its wording concerns independently directing autonomous weapons where applicable law, regulation, or Department policy requires human control. OpenAI cites Department of Defense Directive 3000.09, dated January 25, 2023, which requires rigorous verification, validation, and testing for AI used in autonomous and semi-autonomous systems before deployment.
That distinction leaves important questions unanswered by the public material: whether the restriction covers target recommendations or battle-management support, what counts as meaningful human control, and how a cloud model’s output might feed a separate system that takes action. OpenAI argues that cloud-only deployment prevents its models from directly powering fully autonomous weapons at the edge. That is the company’s stated rationale, not an independently audited technical finding.
Best Value
- Build your own awesome, wearable mechanical hand that you operate with your own fingers.
- No motors, no batteries — just the power of air pressure, water, and your own hands!
- Hydraulic pistons enable the mechanical fingers to open and close and grip objects with enough force to lift them. Every finger joint can be adjusted to different angles for precision movement.
- Three configurations: right hand, left hand, and claw-like; adjustable to fit virtually any human hand.
- Learn how pneumatic and hydraulic systems are used in industrial robots such as automobile components..2021 The Toy Association's STEAM Toy Of The Year Winner
Domestic surveillance has a defined but bounded scope
The updated language addresses intentional domestic surveillance of U.S. persons and nationals, including use of certain commercially acquired personal or identifiable information. It should not be read as a public guarantee against every surveillance-related use. The disclosed material does not explain how the agreement handles incidental collection, foreign intelligence activity involving information about U.S. persons, pattern analysis, or use by another agency or contractor under a separate arrangement.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Why Kalinowski’s resignation matters
The resignation came amid a broader dispute over whether AI companies should accept Pentagon work and how their restrictions should be defined and enforced. Contemporary reporting described Anthropic as refusing certain Pentagon demands concerning surveillance and fully autonomous weapons, while OpenAI said it reached an agreement that retained its red lines. Kalinowski’s departure does not establish that she was involved in Anthropic’s negotiations or that her view represented an OpenAI-wide revolt.
The disagreement highlights a governance fault line: companies may rely on policy, contracts, technical architecture, or a combination of all three, while employees may differ over whether those controls are precise and deliberative enough. “Human in the loop” does not by itself explain how much authority a human has, and cloud deployment may reduce some risks without resolving every question about downstream use. Those are unresolved implementation questions, not proof that the agreement has failed.
What is not established
- The full contract text and independent verification of its restrictions are not established by OpenAI’s public explanation.
- The available sources do not show that the Pentagon agreement specifically covers OpenAI-manufactured robots or robotics hardware; they describe broader AI deployment in classified environments.
- Kalinowski’s resignation alone does not establish that OpenAI’s robotics program changed, that other employees objected, or that the company’s robotics work ended.
- OpenAI’s March 8 statement that it had no plans to replace her was time-specific; the cited reporting does not establish later staffing decisions.
Sources: Engadget’s report on Kalinowski’s resignation; OpenAI’s agreement announcement and March 2 update; OpenAI’s announcement about ChatGPT on GenAI.mil.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
