Figure AI is racing to solve one of robotics' thorniest problems: making humanoid robots safe enough to work autonomously in homes without human supervision. In a recent interview, CEO Brett Adcock outlined the multi-layered safety strategy his company is developing to move beyond lab demonstrations and into real family environments, where a single mistake could cause serious harm .

What Makes Home Robotics Fundamentally Different From Factory Robots?

The challenge facing Figure AI goes far beyond building a robot that can walk and pick up objects. Adcock explained that deploying humanoid robots in homes requires solving problems that factory robots never encounter. A robot working in a warehouse operates in a controlled, predictable environment. A home is chaotic, unpredictable, and filled with hazards designed for human interaction, not mechanical systems .

Consider the everyday tasks Adcock mentioned: doing dishes, folding laundry, and navigating around household items. These activities require the robot to understand context in ways that current AI systems struggle with. A boiling pot of water, a lit candle, or a small child playing on the floor all represent potential disaster scenarios that the robot must recognize and avoid without human intervention .

"We have to get the product cheap enough, we have to make enough of them, we have to make the performance work in very complicated things like walk around a house and do dishes, laundry, very complex things. Small kids can't do this. It takes adults to do this level of work, and we need all that done in a mechanical system that doesn't have any humans around for most of it, that does it autonomously without making any mistakes," explained Brett Adcock, CEO of Figure AI.

Brett Adcock, CEO of Figure AI

How Is Figure AI Building Safety Into Its Robots?

Adcock outlined a two-pronged safety approach that combines hardware design with artificial intelligence. The strategy includes both intrinsic safety, which focuses on the physical design of the robot itself, and semantic safety, which involves teaching the robot to understand and avoid hazardous situations through machine learning .

• Intrinsic Hardware Safety: The robot's physical design prioritizes safety around humans, ensuring that even if something goes wrong, the mechanical system itself won't cause injury. This includes considerations like force limits, material choices, and movement constraints.
• Semantic Safety Through AI: The robot uses neural networks to understand its environment and recognize hazards. This means teaching it to identify boiling water, open flames, and other dangers, then programming it to maintain safe distances and avoid interactions that could cause harm.
• Environmental Awareness: Beyond recognizing specific hazards, the robot must understand the broader context of a home environment, including the presence of children, pets, and fragile objects that require careful handling.

Adcock emphasized that this work is still ongoing. Figure AI has been testing robots in his own home for several months, sometimes with his children present, but he stressed that the company is not yet at the point where he would feel comfortable leaving a robot unsupervised with his kids for extended periods .

What's the Timeline for Trustworthy Home Robots?

When asked about his confidence level, Adcock was candid about the remaining work. He noted that his children have grown comfortable around the test robots, even giving them emotional names and developing attachments to specific units. However, comfort is not the same as safety certification .

"There's still a lot of work to get this thing to a point where we trust it to be autonomous next to my kids all day long in my house. That's the kind of trust you need for one in your house," stated Adcock.

Brett Adcock, CEO of Figure AI

Adcock projected that Figure AI will reach this level of trustworthiness within several years. The timeline reflects the complexity of the problem: the company must simultaneously advance the robot's physical capabilities, improve its AI decision-making, conduct extensive real-world testing, and develop safety standards that don't yet exist in the industry .

The stakes are extraordinarily high. Unlike software bugs that can be patched remotely, a mistake by a physical robot in a home could result in injury or property damage. This reality shapes every decision Figure AI makes about how to test, validate, and eventually deploy its humanoid robots.

Steps to Prepare for Home Robotics Deployment

While Figure AI works on the technical challenges, the broader robotics industry and regulators are beginning to think about how homes and families should prepare for this transition. Based on Adcock's insights, several preparatory steps are emerging:

• Safety Standards Development: The industry needs to establish clear safety benchmarks and testing protocols for home robots, similar to how automotive safety standards evolved. These standards should address both hardware safety and AI decision-making in unpredictable environments.
• Gradual Integration Testing: Rather than deploying robots fully autonomously from day one, manufacturers should plan for phased rollouts where robots work alongside humans, with supervision gradually reduced as confidence builds and performance data accumulates.
• Liability and Insurance Frameworks: Legal and insurance structures need to be developed to address responsibility when something goes wrong. This includes questions about whether liability falls on the manufacturer, the homeowner, or the AI system itself.

Adcock's vision extends beyond safety concerns. He believes that in the coming decades, humanoid robots will become as ubiquitous as smartphones or cars, with most people owning at least one. But reaching that future requires solving the safety puzzle first, which is why Figure AI is taking a methodical approach despite the competitive pressure from other robotics companies .

The company's willingness to test robots in real homes, including around children, demonstrates confidence in its safety approach while also highlighting how much validation work remains. Every month a robot spends in Adcock's home gathering data about how to safely navigate household hazards brings the industry closer to the moment when humanoid robots can be trusted as autonomous household helpers.