What if the world suddenly lost all its wealth? It’s an intriguing idea, but rather than being the plot of the latest futuristic thriller, it’s an idea dreamed up by the What-If Machine, an artificial intelligence (AI) system that is learning to do that very human of activities – think creatively.
AI research has become so advanced in the last five years that researchers can think seriously about using it to create new and unexpected ideas.
They hope creative computers could be used to help inspire novelists suffering from writers’ block, to give children ideas which they could write stories about, or even to individualise mass-produced items. However, they are having to stretch the technology to its limits to get there.
‘We end up pushing the boundaries of artificial intelligence,’ explained Professor Simon Colton, coordinator of the What-If Machine project at Goldsmiths College, University of London.
The machine, which is accessible online, takes information gleaned from the internet and uses complicated algorithms to turn them into ‘What-Ifs’ that combine concepts in unusual ways to create new ideas. For example: ‘What if there was a cloud that had bars instead of water? So rather than being used for bringing a shade, it could be used for buying a beer.’
It filters its results by sending them to project partners at Complutense University of Madrid, Spain, who try to work out whether the What-Ifs have mileage by generating creative stories on the back of them.
‘If the computer doesn’t know what the poem means, then the computer is not being creative.’
Dr Tony Veale, University College Dublin, Ireland
Once the What-Ifs have been scored in this way, they are then sent to a team at the Jožef Stefan Institute in Ljubljana, Slovenia, who use crowdsourcing and machine learning to predict if people will actually like them.
The goal is to hone the system so that it can produce popular ideas more frequently. While they’ve only just started, the early signs are encouraging.
‘We have got statistically significant results indicating that the storytelling approach may work at full scale,’ said Prof. Colton.
The EU-funded project is planning large-scale testing by getting thousands of people to rate thousands of What-Ifs using crowdsourcing.
It’s a similar technique to that used by AI Twitterbots which judge their success on whether or not they are retweeted, such as @pentametron, which pairs and retweets posts that are in iambic pentameter.
‘Sometimes the results can be deeply surprising,’ said Dr Tony Veale, coordinator of the EU-funded PROSECCO project, which is working to develop the field of computational creativity by linking up researchers and communities.
However, these are examples of what he calls ‘mere generation’ rather than real creativity. The AI system uses pre-programmed rules to create seemingly meaningful content which is just that, seemingly meaningful.
‘You get a lot of computers that generate stuff that they themselves don’t understand,’ said Dr Veale, who is based at University College Dublin, Ireland. ‘They generate poetry … but if the computer doesn’t know what the poem means, then the computer is not being creative.’
His group’s Twitterbot @metaphormagnet, which is also part of the WHIM project, uses a store of knowledge to create tweets that it believes to be meaningful. For example, it tweeted in July: ‘Remember when peace was encouraged by nonviolent peacemakers? Now, peace is a victory enjoyed only by conquering victors.’
People judge poetry and artwork by looking at the context in which it was produced and having some background knowledge about the artist.
‘You can fill in a background about him or her as a starving, poor artist in a garret in Paris painting for his or her lunch,’ said Prof. Colton. ‘Whereas, with software, the natural reaction is to think of it as a very cold algorithmic random approach.’
It’s something he has tried to address in his AI-based art machine called The Painting Fool.
The Painting Fool explains the creative process it used after it has finished work. Video courtesy of Simon Colton, who is also the ERA Chair in Digital Games Technology at Falmouth University, UK, based on The Painting Fool’s work
His goal is to grow The Painting Fool into an artist in its own right.
‘When I die, no-one will notice because The Painting Fool will just carry on producing interesting artworks,’ he said.
Now he is considering using similar techniques to augment the What-If Machine.
‘Maybe we’ll regroup and say, “Ok, it’d be great to have a new creative personality in the world, so we’ll build the What-If Machine up to be like The Painting Fool.”,’ he said.
Work like this is laying the foundations of a field that will have a huge impact on our lives in years to come, whether it is in the form of a smartphone app which can help people to think creatively about how to use their time, or as a creativity module that can turn each mass-produced item into an individual, unique design.
‘There is a tremendous fear at the moment about what AI is being used to do, or will be used to do,’ Dr Veale said. ‘Creativity is a wonderful counterbalance to that.’
The WHIM project is one of four that received funding under a specific EU initiative entitled Creative ICT.
The funding, the result of the EU’s Future and Emerging Technologies programme, which seeks to fund new sectors, means that Europe has a leading role in this exciting emerging field.
For more information: http://cordis.europa.eu/fp7/ict/fet-proactive/creativity_en.html
Imagine controlling your computer just by thinking. It sounds far-out, but real advances are happening on these so-called brain-computer interfaces. More researchers and companies are moving into the area. Yet major challenges remain, from user training to the reality of invasive brain implant procedures.
Artificial intelligence is growing ever more powerful and entering people’s daily lives, yet often we don’t know what goes on inside these systems. Their non-transparency could fuel practical problems, or even racism, which is why researchers increasingly want to open this ‘black box’ and make AI explainable.
Stone and concrete structures with the ability to heal themselves in a similar way to living organisms when damaged could help to make buildings safer and last longer.
Artificial intelligence (AI) used by governments and the corporate sector to detect and extinguish online extreme speech often misses important cultural nuance, but bringing in independent factcheckers as intermediaries could help step up the fight against online vitriol, according to Sahana Udupa, professor of media anthropology at Ludwig Maximilian University of Munich, Germany.
Bacteria can give structures an ‘in-built immune system’ to help them last longer.
Independent factcheckers can bring context to AI tools, says media anthropologist.
Dr Kate Rychert studies ocean plate structures.