The Future Through Artificial EyesThe Future Through Artificial Eyes

Technological and social developments constantly influence the way we look at a topic such as artificial intelligence (AI). The myth of the independently thinking machine has been around since the 1800s, both as a harbinger of an ideal world in which people no longer need to work, and as part of a doomsday scenario in which society loses control of the robots. AI is the ability of a computer to process information and make decisions. Traditionally, it would do so based on instructions and rules set by programmers: if this, then that. Over the last 20 years, however, this has started to change. Instead of explicitly telling a computer what to do, so-called neural networks, a type of algorithm modelled on the human brain, are given a task without clear instructions on how to solve it. By letting the system try different strategies again and again, it can improve itself until it reaches a high rate of succes. While the concept of a self-learning algorithm has been around for decades, the exponential growth in processing power and the increasing availability of large amounts of data over the past 20 years have transformed it from an academic concept to an everyday reality.

The self-learning algorithm has become almost synonymous with AI, yet its workings and implications remain a mystery to most.

In this multimedia installation by Richard Vijgen, a computer vision neural network searches for patterns in 20 years of broadcasts of the future-themed V__VPRO Tegenlight_ _TV show. Visitors can influence the outcome by playing with the way the 'artificial eye' views the broadcasts. The installation shows, as it were, an archive of visions of the future screened during 20 years of VPRO Tegenlight.

Surrounding the multimedia installation are entry points allowing visitors to unpack and demystify processes that are essential to computer vision and neural networks. This will give further insight and clarity into what is commonly bundled into the single term: artificial Intelligence. Visitors can dig deeper into the technologies by analysing how algorithms 'see' patterns in images or texts; explore how biases are at the foundation of those same algorithms; challenge ideas of the 'intelligence' of these systems and see how the choice of training data effects the output of artificial cognitive systems. All of this is supported by interviews from previous episodes of VPRO Tegenlight on subjects like computation, dataism and civil rights.

Het Nieuwe Instituut is excited to celebrate 20 years of the futuristic vision of _VPRO Tegenlicht! _This documentary-style TV programme, from broadcaster VPRO, examines the national and international developments that are shaping our world in the 21st century. The programme's signature is to be present where new ideas are developed, tested and criticised in the worlds of politics, economics, society, technology and science. You can find _VPRO Tegenlicht _on TV, online, on social media, at local Meet-ups, by listening to its podcast and and reading its weekly newsletter.

With the generous support of the Mondriaan Fund, this exhibition and the installation by Richard Vijgen are part of the larger 20-year celebrations. There will aslo be a VPRO Tegenlicht book designed by Irma Boom featuring newly written essays and international interviews. Together with Sudox, VPRO has also developed the web platform Archive of the Future. Using algorithms for image and person recognition, and speech and text analysis, all VPRO Tegenlicht broadcasts since 2002 have been transformed into a time-coded dataset, which allows users to search the collection for fragments, quotes and even individual shots. The timeline- and theme-based interface allows for cross-referencing and is intended for scholars, teachers, researchers, students, artists, designers and anyone else wanting to explore stories, fragments, images and quotes from the Tegenlicht archive.