'World's most advanced' humanoid robot does eerily accurate impressions of Elon Musk and Donald Trump

In the fast-paced era of artificial intelligence (AI), it feels like every week brings an interesting new development.

Now it's the turn of Ameca, a robot made to look like a humanoid torso and head by UK-based firm Engineered Arts.

A recent video showed off what Ameca was capable of - the robot is first asked to describe the room around it and to identify various objects placed in front of it. It does so really well, including identifying "an anatomical model of a human head, quite a detailed one".

The robot seems to be able to riff further than just pure identification, too, saying things like: "Fascinating, isn't it, how the organic is replicated for study?" This sounds closer to real conversation than most other examples we've seen, with prompts going back and forth rather than just in one direction.

The video gets a lot juicer from there, though, as the demonstrator, Engineered Arts founder and CEO Will Jackson, asks the robot to mimic a cast of famous people.

First up, we get Morgan Freeman, and while we've heard better impressions in our life, it's still decently close, albeit a little deeper than it should be.

Next up, the robot is asked to tell a story about the exploration of Mars using Elon Musk's voice, and the impression it offers up is far better. It tells of "robot rocket ships, each one smarter than the last" jetting up to the red planet - which sounds like something Musk himself would talk about.

Finally comes the best bit of the video: when the robot is asked to do a Donald Trump-style rant, in the voice of Spongebob Squarepants - and it does a terrific job.

Engineered Arts / YouTube

It's quite bizarre to hear a Trumpism come from the robot such as: "We're going to discover things, big things - and let me tell you, nobody loves Mars more than me. It's true, we're making space exploration great again".

Ameca is far from a finished platform right now, as Engineered Arts happily admits - these are early days for this sort of system. It's envisioned as a 'platform for human-robot interaction', according to the company, meaning it could eventually be a way to talk to complicated systems in a more understandable way.

For now, though, without the ability to move around and with quite a lot of scripting still required to apparently make some of its demonstrations work (it's unclear if any script was written for this video, either), you don't have to worry about Ameca taking over your job any time soon.