Robot says AI is ‘threat and opportunity’ as owner forced to reset it during House of Lords address

The robot, named Ai-Da, uses algorithms to write poetry and create artworks

A ‘realistic’ robot has said that artificial intelligence could be both a “threat and opportunity” to artists, in the first address by a robot to the House of Lords today.

The robot, named Ai-Da after the 19th-century mathematician Ada Lovelace, gave evidence to the House of Lords Communications and Digital Committee as part of an inquiry into the future of the arts, design, fashion and music industries and how artificial intelligence might affect them.

While the robot is providing evidence, it is not a witness in its own right and does not occupy the same status as a human. In one instance, Ai-Da turned off while answering questions, leading its operator to have to reset the machine.

“I produce my paintings by cameras in my eyes, my AI algorithm, and my robotic arm ... for my poetry, using neutral networks, this includes analysing a large corpus of text ... and using the structures and contents to generate new poems. How this differs from humans is consciousness. I do not have subjective experiences, despite being able to talk about them”, Ai-Da said. “Although not alive, I can still create art.”

Since the robot has been trained on language made by humans, it is understandable that it would claim to be alive - as humans are unlikely to describe themselves as dead when speaking or discussing art.

The robot has different algorithms for creating artworks compared to creating poetry.

“Technology has already had a huge impact on the way we create and consume art, for example, the camera and the advent of photography and film, and it is likely this trend will continue”, she said. “There is no clear answer as to the impact on the wider field, as technology can be both a threat and an opportunity for artists creating art”.

Ai-Da, which was devised in Oxford by Aidan Meller, a specialist in modern and contemporary art, before being built in Cornwall by Engineered Arts, had previously painted a portrait of the Elizabeth II ahead of the Platinum Jubilee. The robot has cameras in its eyes and uses computer algorithms to process human features.

“Ai-Da uses an AI language model”, Mr Meller said, adding that questions were provided to the team beforehand. While Ai-Da can respond in real-time, the answers would not be as high-quality, he continued. Mr Meller remains liable for anything the robot says.

“When it comes to traditional art practises, some say that it cannot be used”, the robot said, “but others say it can be valuable. Because art is often up to interpretation, the role of the audience is key. In general, contemporary artists often seek to engage with their audience and create a dialogue about important issue and ideas. As long as technology remains an important issue in society, there are not many limits to how it can be used.”

In general, while artificial intelligence was alleviating some artists from repetitive work, such as writing radio jingles which are now almost entirely synthesized, but said that there were serious risks on the labour market, Paul Fleming, general secretary of Equity, another witness at the hearing, said.

Radio jingles, for example, paid well because the work was repetitive and high wages was necessary to invigorate the artist, he said, but artificial intelligence could see that money disappear from musicians’ pay packets - and continues as companies could use the technology to remove them from TV and film work.

Robots could also be used to create books, and while would not be able to “produce tearjerkers” at the height of human literature, Dan Conway, chief executive officer of the Publishers Association, said, could produce “lots and lots of work that is low quality” that could push out human writers by sheer quantity.

Currently, a robot cannot claim copyright of a piece of work because of regulation around what constitutes ‘originality’, which is based on human characteristics that a robot does not possess.

“At the moment, we don’t have case law”, said Dr Andres Guadamuz, a reader in Intellectual Property Law at the University of Sussex. Moreover, while many copyrighted works by humans are used to train artificial intelligence that has been allowed by legislation as long as it is in the name of research, rather than commercial use.

There will be new spheres of art that are opened up by AI, the witnesses said, but there was an issue in guaranteeing protection for human workers.

“AI is responding to markets. It is producing content for consumers that already exist. The real question for the creative industries is: do we have a sufficient framework to be adequately subsidising human content in order to intervene in that market?” Mr Conway said.