Boston Dynamics has turned its robot dog Spot into a talking tour guide that is able to take on different ‘personalities’ according to the situation.
The US robotics company gave its four-legged robot the ability to speak using OpenAI’s ChatGPT API, as well as a number of other open-source large language models (LLMs).
In a blog post, the company said that it was interested in using Foundation Models – large AI systems trained on massive datasets – as autonomy tools, allowing its robots to make decisions in real time based on the system’s output.
In particular, the company is interested in the ability of LLMs such as ChatGPT to “roleplay, replicate culture and nuance, form plans, and maintain coherence over time”.
This was paired with recent Visual Question Answering (VQA) models that are able to understand the content of an image and answer questions about it.
The result is a robot tour guide that is able to walk around, capture images of objects and items in the environment, describe them, and respond to questions from the audience.
Robot adopted different personalities depending on the subject
Spot was given a brief script for each room in the tour of the Boston Dynamics facilities, which it combined with VQA capabilities to provide relevant information and answer questions.
It also assumed different ‘personalities’ for the different areas, taking on different accents and language use, along with visual embellishments.
It even moves its mechanical gripper like a puppet’s mouth as it speaks from an equipped speaker.
A video posted by the company showcased a ‘fancy butler’ persona that spoke in a British accent while sporting a moustache and top hat.
Other personalities include a ‘teenage robot’, whose speech is peppered with ‘like’ and ‘totally’.
Another is based on an Indiana Jones-style 1920s archaeologist, which talks about its adventures and hat, along with the actual tour.
There’s a ‘Shakespearean time traveller’ that speaks in rhyming couplets and a sarcastic, nihilistic personality that the company’s software engineers have simply dubbed Josh.
At one point in the video, Josh describes what he sees as “the unfathomable void of my existence reflected in this QR code-filled board”.
“Oh, and also a large window,” he adds, before going on to compose a gloomy haiku about a room with a humming generator.
Today’s news was brought to you by TD SYNNEX – the UK’s number one solutions distributor.
Read more of our latest Industry Updates stories
