Is Ameca autonomous?

No, its behavior is scripted, LLM-driven, or teleoperated

No, not in the way viral clips suggest. Ameca's conversation is scripted or LLM-driven and its facial expressions are pre-programmed, and it is sometimes teleoperated. It is not sentient, self-aware, or independently acting.

Key facts

Form factor: humanoid
Maturity: commercial
Verified deployments: 2
Maker: Engineered Arts
dof: 61
specs: Ameca: 61 actuated movements (27 DoF head/face + 34 DoF upper body), 50+ facial expressions, ~187 cm, stationary; Tritium software with LLM/speech integration; Generation 3 shown at ICRA 2025. ~$250,000; offered for purchase AND rental.
ai system: Tritium
height cm: ~187 cm
formFactor: humanoid (stationary expressive upper-body/face HRI humanoid; does NOT walk)

Engineered Arts on the registry →

The short answer

No. Ameca is famous for videos that look spontaneous and self-aware, but that is a demonstration effect, not autonomy. Its behavior is scripted or driven by a large language model, its facial expressions are pre-programmed, and it is sometimes teleoperated by a human.

Separating the demo from the claim

This is where careful reading pays off. Engineered Arts built Ameca as an expressive human-robot-interaction platform. When Ameca appears to converse, its speech may be generated by an LLM, but the more than 50 facial expressions are pre-programmed sequences, and in many public demonstrations a human is teleoperating it. The registry's media note is explicit: Ameca is not sentient or self-aware. There is no autonomous decision-making about the physical world here, and Ameca does not even move through space, since it does not walk.

What is verified

What is real: a genuinely advanced expressive face, LLM and speech integration via the company's Tritium software, and commercial installs at named museums and institutions. What is not established: any claim that Ameca acts, thinks, or decides on its own. Treat the lifelike moments as choreographed capability, scripted, LLM-driven, or teleoperated, rather than autonomy.

For how DEPLOY separates a scripted demo from a verified autonomous capability, see verified-vs-claimed.

Frequently asked

Is Ameca sentient or self-aware?

No. The registry states plainly that Ameca is not sentient or self-aware. Its expressions are pre-programmed and it is sometimes teleoperated.

Does Ameca think for itself?

No. Conversation may be LLM-generated, but facial behavior is scripted and a human often teleoperates it. There is no autonomous decision-making about the physical world.

Verified vs claimed

Verified

Verified: Engineered Arts (founded Oct 2004, Falmouth, Cornwall UK; founder Will Jackson) makes Ameca, an expressive stationary upper-body/face social humanoid for human-robot interaction, research, exhibition, and hospitality. First revealed Dec 2021, public debut CES Jan 2022, Gen 3 at ICRA 2025. Predecessor RoboThespian has a 50+ unit track record (NASA Kennedy, science museums).

Claimed / roadmap

Claimed but NOT verified: Exact total Ameca install count (the '50+' applies to RoboThespian, not Ameca); precise per-config pricing (only ~$250k has a named-quote source; the $100k-$500k range is secondary-blog); 'world's most advanced humanoid' (marketing superlative); future walking capability (stated in-development, not demonstrated).

Split from the registry record. See the verified-vs-claimed framework.

Where it is deployed

2 verified deployments on the registry for the Ameca. Click any marker to open its primary-source record.

Explore the deployment map

See where these robots are verified operating, by place and type.

Every robot coming to market, tracked and verified.

DEPLOY is the registry of physical AI: every company, model, deployment, and safety record, with the sources behind each one. Browse the humanoids, robotaxis, drones, and delivery robots being deployed right now.

Explore the DEPLOY registry →

Most-tracked robots