For autonomous robots in hospitals, functional design is not enough. To avoid alarming vulnerable patients, robots must be intentionally designed to appear non-threatening and trustable, using visual cues to acknowledge people and signal benign intent.

A key to human-robot interaction is managing expectations. A robot that suddenly turns is alarming. However, if the robot first looks in the direction it intends to move and then turns, it signals its intent, making the action feel natural and non-threatening to humans.

Figure is intentionally designing its robots to avoid two extremes: menacing appearances and overly friendly looks with "googly eyes." The goal is to position the humanoid as a sophisticated, high-end piece of technology—a tool for humanity—rather than trying to fool users into thinking it's a toy or a person.

The humanoid form factor presents significant safety hazards in a home, such as a heavy robot becoming a “ballistic missile” if it falls down stairs. Simpler, specialized, low-mass designs are far more cost-effective and safer for domestic environments.

Jony Ive added a handle to the iMac not for portability, but to make intimidating technology feel friendly. By encouraging users to touch the machine, it created an emotional relationship and gave them 'permission' to interact, overcoming the fear many felt towards computers at the time.

Unlike fixed industrial robots, a simple emergency power-off is unsafe for humanoids. They require constant energy to balance, so an emergency stop would cause them to fall over, creating a new and unpredictable hazard. This fundamental difference requires an entirely new set of safety protocols for the industry.

Standard Bots CEO Evan Beard argues that a key barrier for domestic humanoid robots is safety and robustness, which he calls the "Home Alone test." A robot must be able to withstand unpredictable, chaotic interactions, like children jumping on it, a scenario that current RL training methods cannot adequately simulate or solve.

Widespread acts of public aggression and vandalism against food delivery robots are more than just pranks; they are an outward expression of a deeper public anxiety, particularly in America, about the intrusion of AI and automation into daily life. This physical backlash suggests significant societal unease with technology's encroachment on human spaces and jobs.

The user's experience with Clawdbot produced two conflicting feelings: 'this is so scary... nobody should be doing this' and 'boy, oh boy, I want this thing.' This emotional dichotomy captures the current state of agentic AI, where the desire for its power is in direct conflict with its profound risks.