At the superficial level, a robot at a tradeshow may seem like a gimmick designed to draw attention to itself and bring a crowd to a booth. Measured that way, Ameca did its job for its designer, Engineered Arts. It was quite a sensation at CES2022 attracting throngs to Engineered Arts’ booth, along with numerous media mentions.

Ameca is another step on a 15-year journey for Engineered Arts, which started with a group of artists and engineers designing and building mechanical actors for the Mechanical Theater at the Eden Project. This led to the RoboThespian Mark 1. Over time, Engineered Arts developed a series of robot siblings including

• Ameca (“The future face of robots”),
• Mesmer (“Your human lookalike”), and
• Quinn (“Your robot customer assistant”).

Notable customers include Amazon Prime Video, National Geographic, and Madame Tussauds of London. The applications go beyond novelty and Ameca is seeing service as a customer service robot as well as educational settings (e.g., museum deployments, science centers, and universities). The business model is both rental and for sale.

Not a Wizard, but a TinMan

While it is a commercial project, it is also a research platform as Roe points out. Their robot operating system, Tritium, which is a combination of hardware, software and cloud implementation works across their robots and allows for customization by its customers.

To create real-time interactions, they have a telepresence feature, TinMan. This allows conversational artificial intelligence (natural language processing) to be a sort of first line, whereby the computer/robot answers the easy questions (e.g., checking guests into a hotel).

Like a real-world chatbot, the conversation would transition to a human-driven one for those questions beyond the ken of the robot. The difference is that the robot continues to be the eyes, ears, and voice, so the human (at least this one) does not detect the transition between human and robot interaction.

As with other telepresence applications, Engineered Arts envisions one human managing a number of robots, effectively leveraging the human’s ability to, say, provide directions at a large facility, for instance.

Although Roe says it is too early to provide movement to their robots, it isn’t too difficult to imagine them someday sitting on autonomous chairs, such as profiled here, providing mobile assistance and monitoring public spaces.

A Humbling Conversation with Ameca

Interviewing a robot that is more articulate than oneself is humbling to say the least. Simply, Ameca does not use filler words like this author (edited out in the interview below to reduce the interview time to a minute). Its response time is amazing (no editing of its portion of the conversation; that is its actual response time to questions which were made up on the spot).

Ameca shows its class in the way it responded to my comment, which HR might deem inappropriate, about its beautiful eyes (which are also cameras). Finally, Ameca’s manners are better than yours truly, as it said goodbye as I walked away from it. What is amazing is its facial expression and how it tracks one’s face (see the above video for how it looks towards Morgan when he says “Ameca”).

What was not obvious was that TinMan must have been part of the demonstration and that Ameca was getting an assist from an operator in the background. Still, Ameca served its purpose by not only drawing a crowd but being able to respond to an impromptu interview in an engaging manner.

[added 11/09/22 – In a follow-up email, Roe confirmed that TinMan was the vehicle for Ameca to engage in the conversation below. Roe wrote, “TinMan was used for remote operation. It works in a similar way to a video conference but using a robot. It can work well from around 2mbps and can work from any connection. As it doesn’t walk we tend to use a wired connection for the best reliability.”]