An AI Robot Testifies at the House of Lords

For the first time in British history, a robot testified in the upper chamber of Parliament. The humanoid robot, named Ai-Da, spoke about the future of artificial intelligence in creative industries. Ai-Da made sure to dress up for the occasion. She was wearing a bright orange shirt, denim overalls, and a sleek bob and bangs hairstyle to go with the outfit. She took questions from members of the House of Lords Communications and Digital Committee. The goal of the session was to discuss technology’s role in art. Let’s take a closer look at Ai-Da as well as the data annotation that’s necessary to create conversational robots. 

Who is Ai-DA?

Ai-Da is named after a British mathematician Ada Lovelace who is widely considered to be the first computer programmer. The robot was created around three years ago by Aidan Meller, who is a professor at Oxford University, along with his team. 

The robot artist paints portraits, with subjects ranging from Elizabeth II to Billie Eilish, and its works have been exhibited at the United Nations and the Venice Biennale. Ai-Da also writes poetry using an AI algorithm that processes and synthesizes existing poems to learn about various styles and subjects. To paint, the robot relies on data from AI algorithms, cameras in its eyes, and mechanical arms designed to maneuver a paintbrush.

How Did Ai-Da Do at the Hearing?

According to sources, Ai-Da answered some prepared questions from the lawmakers, making it unclear how the robot would handle itself in a regular conversation. Also, Ai-Da malfunctioned during the hearing, and her creator needed to reset the robot. In case you were wondering, while Ai-Da provided some evidence to the committee, it was not a witness in its own right and did not have the same status as a human. Still, Ai-Da is a very advanced robot that can understand human speech and navigate in its surroundings. Let’s take a look at the data annotation that makes this possible. 

What Data Annotation Methods are Required to Create a Robot Like Ai-Da?

Since Ai-Da can move around and recognize objects in her surroundings, 3D Point Cloud annotation was necessary to train Ai-Da to walk around the physical world just like a human. 3D Point Clouds are created by LiDAR and are a digital representation of how an AI system sees the physical world. The way it works is the LiDAR sends out pulses of light that bounce off objects and return back to the LiDAR. The longer it takes the light to return, the farther away the object is located. 

Data annotators would need to label all of the objects in the 3D Point Cloud with methods ranging from labeling to semantic segmentation. In addition to this, the images may need to be color coded to show their proximity to the LiDAR. For example, objects close to the LiDAR will be in blue since this color has a short wavelength. Farther away objects would be in orange or red since these colors have longer wavelengths. 

Now, let’s take a look at how Ai-Da was trained to understand human speech. First of all, the needed training data needs to be assembled that would cover various scenarios. In Ai-Da’s case, this is simply because the questions were prepared, and all the researchers would need to do is input those questions into the ML algorithms. However, this text would also need to be annotated with techniques for natural language processing (NLP). This includes things like entity annotation, which is where data annotators locate, extract, and tag entities in the text. This includes things like names and various parts of speech. Entity linking would then need to be done to connect all of the entities so the system can learn how all of the various parts of speech work together in a sentence. 

Trust Mind Support With All of Your Data Annotation Needs 

Mindy Support is a global company for data annotation and business process outsourcing, trusted by several Fortune 500 and GAFAM companies, as well as innovative startups. With nine years of experience under our belt and offices and representatives in Cyprus, Poland, Romania, The Netherlands, India, and Ukraine, Mindy Support’s team now stands strong with 2000+ professionals helping companies with their most advanced data annotation challenges.