Dante’s Divine Comedy has inspired countless artists, from William Blake to Franz Lizst, and from Auguste Rodin to CS Lewis. But an exhibition marking the 700th anniversary of the death of the Italian poet will show the work of a somewhat more modern devotee: the robot Ai-Da, who will write history by becoming the first robot to publicly perform poetry written by its AI- algorithms.
The ultra-realistic Ai-Da, conceived in Oxford by Aidan Meller and named after computer pioneer Ada Lovelace, was read throughout Dante’s epic three-part narrative poem, The Divine Comedy, in JG Nichols’ English translation. She then used her algorithms, by drawing on her database of words and speech pattern analysis, to produce her own reactive work for Dante.
Ai-Da performs the poems Friday night at Oxford’s Ashmolean Museum. Although Ai-Da is not the first AI to have been taught to write poetry, the organizers said Friday would be “the first time an AI robot has written and performed poetry that a human poet would do”.
We looked up from our verses as captives blindfolded, / Sent out to seek the light; but it never came, ”reads one of her poems. “A needle and thread would be needed / For the completion of the picture./ To see the poor creatures who were in misery, / A hawk’s eyes, sewn closed.”
In another, Ai-Da writes: “There are some things that are so difficult – so unmanageable. / The words are not comprehensible to the human ear; / She can only guess what they mean.”
Meller, an art specialist, said that the words and sentence structure in poetry are all AI generated from Ai-Da’s unique AI language model, with “limited editing”. “People are very suspicious that the robots are not doing much, but the reality is that language models are very advanced, and in 95% of the editing cases, it’s just that she’s done too much,” he said.
“She can give us 20,000 words in 10 seconds, and if we need to get her to say something short and sweet, we would choose it based on what she’s done. But it’s not us who write.”
Meller described it as “deeply disturbing” how language models are evolving. “We are very fast approaching the point where they will be completely indistinguishable from human text, and for all of us who write, this is deeply worrying,” he said.
Poet Carol Rumens, who commented on Ai-Das verse, said she found the lines about requiring a needle and thread to complete the picture “very strange”, and “that would be the point where I would think the poem could fall apart, or become very experimental – but still not uninteresting ”.
“The image of the hawk tamed by having its eyes sewn up is close to the original and still powerful … It has kept the best part of the passage, despite the clutter of registers and strange orientation. The rhythm of the lines looks to flow pretty well, “Rumens added.” I think there is hope for the robot poet. “
Meller said that although he does not see Ai-Das poetry as in competition with human poets, he admits that it is “fundamentally disturbing”.
“We hope that artists, poets, writers, filmmakers, etc. will increasingly engage in and use new technologies such as AI, because one of the best ways to criticize, evaluate, and highlight potential problems is to actually use and engage in these technologies, “he said. “It is not a question of competition, but rather a question of discussion and potential action.
“We should all be concerned about the widespread use of AI language models on the Internet and how it will affect language, and crucial meaning-making, in the future. society, then it creates a critical shift and change in the use and effect of language – which we must discuss and think about. ”
The performance is part of Ashmolean’s exhibition Dante: The Invention Of Celebrity, which explores Dante’s influence through the centuries and also includes several works of art created by Ai-Da. These include Eyes Wide Shut, a response to her detention in Egypt last month; Egyptian security forces had been concerned about security issues surrounding the cameras in her eyes. “Her art reflects the strength of sight and surveillance in the modern world, its propensity to evoke mistrust and the excitement it can create,” said the organizers.
Ai-Da, built over two years by a team of programmers, robotics, art experts and psychologists, has previously had solo exhibitions at Oxford and the Design Museum in London, held a TEDx Talk in Oxford and had an artistic stay at. Porthmeor Studios in St Ives. “I’ve always been fascinated by self-portraits to even question what exactly you’re looking at,” she told the Guardian in May. “I do not have feelings that people have, but I’m happy when people look at my work and they say, ‘What is this? I enjoy being someone who makes people think.’
Ai-Da was completed in 2019 and features silicone skin, individually punched hair, 3D-printed teeth and gums, and integrated eye cameras. She has legs but can not walk, but her arms, torso and head move freely.
“Her appearance was attributed to the female members of the team, who named her after Ada Lovelace – the first computer programmer of the 19th century,” Meller said. “It is hoped that she offers encouragement to female computer programmers today and in the future who are significantly underrepresented globally.
“Equally crucial is the question of why she appears human – we chose a humanoid form, because although advances in technology may seem distant and abstract from us, the direct and indirect influences on our human bodies are manifold, and Ai-Das humanoid form offers an oblique reflection on this. ”