How creative can Artificial Intelligence be?
Artificial Intelligence in creative professions
Artificial Intelligence has become creative: Thanks to machine learning, computer programs can create music, texts and images. And it is also used in design, at least to a degree.
One thing for starters: this article was not written by a computer. That may sound obvious, but it isn’t. Readers of the Los Angeles Times will certainly know better, because the LA Times already uses robot journalists to write simple, factual reports on earthquakes — in record time. At other news desks, they write up the sports results or prepare financial reports. Artificial Intelligence (AI) expedites creative work.
Intelligent machines of today can do a lot more than just beat people at chess: Watson, the supercomputer from IBM, is capable of diagnosing rare diseases better than any human doctor. But Watson can also create trailers for feature films. And the scientists at the Sony Computer Science Laboratory in Paris are already publishing their first pieces of music composed by software using Flow Machines algorithms. So, as you can see, AI already has a foothold in a variety of creative professions. “Providers like The Grid or Wix are using simplified AI in the field of web design,” explains Olivier Heitz, Senior Product Designer, Artificial Intelligence & Machine Learning Group at Swisscom. “AI recognises which templates are best suited to which application. This enables smaller companies and private individuals to create a website at little cost that is not created from a template, but as a ‘bespoke solution’.”
“If it were possible to leave some of the more mundane work to Artificial Intelligence, that would mean a great step forward.”
The designer develops and AI does the donkey work
Designers can benefit from intelligent software as well. Since the work they do usually focusses on the user, designers perform a lot of research in advance, solicit user feedback and define a clear problem. They then develop different approaches, create prototypes, test them, reject them and specify solutions. As a part of the design process, revisions and iterations are the order of the day: “You frequently get to the point where you have to try out lots of different design variants to be able to move forward,” explains Heitz. “However, unproductive, repetitive work costs you a lot of creative energy. If it were possible to leave some of the more mundane work to Artificial Intelligence, that would mean a great step forward.”
AI can also help with mood boards, where it is mainly used to search for ideas: Designers use mood boards to collect photos, drawings, colours, materials and texts in order to capture the mood and effect of their design concept. AI would certainly be worth considering here. “Theories abound concerning size ratios and the effect that fonts, colours and layouts have,” explains Heitz. “AI could be used to make specific design proposals based on these rules.”
Slime moulds influence the design of aircraft parts
AI has certainly made inroads into the world of industrial design. Autodesk, a software company for digital 2D and 3D product design, has shown this with its recent launch of Project Dreamcatcher. Dreamcatcher is a generative design system, where it is not the designer who produces the desired result, but a programmed algorithm. With Dreamcatcher, the user simply defines the specifications, such as functional requirements, material types, manufacturing methods or cost constraints. The system then runs through a database and develops suggestions that match the specifications. Dreamcatcher uses algorithms based on nature and imitates how certain shapes have been preserved or rejected in the course of evolution. The allows the software to create highly complex shapes which are much more resilient than conventional designs. Dreamcatcher has already been used to design chairs, a chassis for racing cars and partitions for planes. The growth patterns of slime moulds and the structure of mammalian bones were used as inspiration for the latter.
However, Artificial Intelligence has not spread to all design sectors yet. Professional designers of digital products such as website or user interfaces will find hardly any suitable AI tools: “If I’m designing apps or websites, there’s nothing in the AI field that I could not do without,” Heitz explains. In fact, we’ve been using the same programs in design for 15 years, and Adobe Suite remains the standard at most companies with software such as Illustrator or Photoshop.
“It’s up to us what comes next.”
However, this could change in the future. According to Heitz, design schools are placing increased emphasis on students learning these technologies. It is important to have this knowledge to enhance the use of AI in design. “It’s up to us what comes next,” says Heitz. “The more we address and engage with creative minds, the more likely it is that the next tools will be really helpful.”
This article was written by Christoph Widmer and originally appeared on swisscom.ch.
What is Machine Learning?
The principle of machine learning forms the basis of AI as we know it today, with the computer or the software independently generating new knowledge based on experience. The systems analyse data and find patterns, laws or distinctive features using intelligent algorithms. The programmer provides feedback and optimises the system. AI thus becomes more precise with every pass it makes. Based on this experience, AI can ultimately link data, establish connections, draw conclusions or make predictions.
Making AI easy(er) to understand. Check out our AI Glossary.