Under the title "AI: more than human", the immense Barbican cultural centre brings together more than 200 installations, exhibits and projects by artists, scientists and researchers from all over the world. 

From Thursday until August, visitors will be able to take a journey from the long-held dream of creating artificial life to the reality of today's most cutting-edge projects. 

An immersive space by Japanese collective teamLab forms one of the most intriguing exhibits, with art and science combining to let the visitor leave their mark on an evolving digital wall projection. 

There are also robots of all shapes and sizes, from Sony's small dog Aibo -- whose first version from 1999 has now evolved into an AI model -- to a large mechanical arm that prepares and serves cocktails.

Other exhibits explore the complex systems that keep big cities ticking over and push forward research into medical conditions from cancer to blindness. 

The current limits of AI are investigated, including racial bias in some facial recognition software.

Properly designed AI can help prevent harm, Francesca Rossi, head of ethics at IBM Research, told AFP. 

"If the machine can understand this concept of bias, then it can alert us if it sees that there is bias in our decision making," she said. 

Inserting human values

Although the idea of decoding the human brain and imitating its functions was born in the mid-1950s, AI only exploded in 2010 thanks to very fast state-of-the-art processors that allow the analysis of huge amounts of data. 

The machines have since come on leaps and bounds.

IBM's Deep Blue beat Russian chess champion Garry Kasparov in 1997 while AlphaGo -- developed by Google's DeepMind team -- in 2016 beat Lee Sedol, world champion in the 3,000-year-old Chinese board game known as Go. 

Both are present in the exhibition, helping to outline how AI can help solve problems of enormous complexity, such as climate change. 

"The thing that we dream about would be, what if a machine could say: 'here is a clever way of changing how we run our economy that fixes climate'," explained Swedish philosopher Anders Sandberg, Senior Research Fellow at the Future of Humanity Institute, Oxford. 

But for that, "we need to find a good way of putting human values into machines so they will act without accidentally harming you", he added, joking that AI could conclude the best solution was to eradicate human beings. 

Despite its ambitious scope, the exhibition is only one part of a larger project called "Life Rewired", which explores the impact of technology on society. 

The centre recently held a concert of baroque music composed by AI after analysing works by Johann Sebastian Bach. 

"We gave the machine-learning algorithm all of Bach's keyboard works," explained the project's architect Marcus du Sautoy.

"That's a lot of music, but often machine-learning needs millions of data points to learn from", added the Oxford mathematician.

He hopes to demonstrate that, rather than competing with humans, artificial intelligence can help humans "to think outside of our narrow creative window".

"Humans get very stuck in ways of thinking, we often end up behaving like machines," he said.