7# From Darwin to DeepMind

Artificial Intelligence and the Age of the Superhuman

Long ago, humanity lived in paradise, in a state of ignorance. When Eve offered Adam the apple from the Tree of Knowledge and they ate from it, they gained access to knowledge that brought them closer to the natural process of evolution.

The death that awaited them was not a punishment but part of the natural order, which Darwin described through the principle of the "survival of the fittest".

The Tree of Life, from which they were kept away, symbolised not immortality but the continuous process of adaptation and evolution.

As a famous saying often attributed to Darwin goes: "It is not the strongest of the species that survives, but the one most responsive to change."

Humanity entered the process of evolution not through divine permanence but through the ability to adapt continuously.

The latest developments in artificial intelligence suggest that science fiction may soon become reality.

Machines are learning with increasing autonomy and becoming more and more human-like.

A computer may consume as little as 20 watts of power, while the human brain accounts for approximately 20 per cent of the body's total energy consumption.

Computers already seem to surpass humans in many areas.

The Superhuman Experiment Has Begun

Since the late 1990s, roboticists, neuroscientists and psychologists have been attempting to transfer human cognitive abilities to computers and robots.

Their goal is to create a self-learning machine capable of thinking as well as, or even better than, humans through mathematics and algorithms.

Meta-learning is an approach that enables machines to learn how to learn. Artificial neural networks, meanwhile, are designed to imitate certain aspects of how neurons in the brain process information.

Such networks form an important foundation of artificial intelligence. They consist of numerous interconnected computational units, often called artificial neurons, arranged in successive layers.

In a typical machine-learning process, a computer is shown an enormous collection of images, such as photographs of dogs and cats, without programmers having to specify every visual feature that distinguishes them.

Over time, the system identifies patterns and learns to recognise the differences between images.

With each training cycle, the network adjusts the connections between its artificial neurons, strengthening patterns that contribute to more accurate results.

Computers can also be trained using videos, speech recordings and music.

In this way, the potential of artificial neural networks is increasingly developed, allowing machines to perform a growing range of tasks.

"The Development of Full Artificial Intelligence Could Spell the End of the Human Race"

— Stephen Hawking

One of the pioneers in artificial intelligence is DeepMind, a British company that has made significant advances in machine learning.

Founded in September 2010 as DeepMind Technologies by Demis Hassabis and two co-founders, the company was acquired by Google just four years later in a deal reportedly worth approximately half a billion US dollars.

The investment was intended to open the door to extraordinary technological possibilities.

Among DeepMind's early investors was Elon Musk, the entrepreneur associated with Tesla, who has repeatedly expressed concerns about the existential risks of artificial intelligence.

Like Stephen Hawking, Musk has warned that advanced AI could pose a threat to humanity.

The Israeli-American computer scientist Judea Pearl, a professor at the University of California, has also raised questions about the behaviour and decision-making capabilities of intelligent machines.

The underlying concern is whether machines that operate with increasing independence might eventually behave in ways that conflict with human intentions.

DeepMind Wants to Discover New Possibilities

Researchers want to understand human intelligence because the brain is extraordinarily complex, and neural networks can differ greatly in their structure and organisation.

These networks form connections in different ways depending on their architecture and learning processes.

DeepMind's research seeks to understand the principles underlying intelligence and apply those insights to artificial systems.

The ultimate ambition is to develop machines with increasingly general cognitive abilities – perhaps one day creating something resembling a "superhuman" intelligence.

Machines are already becoming part of everyday human life.

Surveillance cameras, industrial robots assembling components, Alexa playing our favourite playlists and Siri telling us tomorrow's weather are all examples of technology becoming integrated into our routines.

The human mind serves as an inspiration for developing universal, self-learning computer systems.

A central function of these systems is to analyse data and learn independently from experience.

With access to immense quantities of information, artificial intelligence could open the door to a future of seemingly limitless possibilities.

Some futurists have predicted that by 2045, artificial intelligence could reach a level capable of matching or surpassing human cognitive abilities.

"The union of human and machine will allow us to create a powerful instrument and multiply our inner potential," argues futurist Ray Kurzweil.

Born in 1948, Kurzweil became known for innovations such as the flatbed scanner and reading machines for blind people.

A pioneer in artificial intelligence and a former director of engineering at Google, he has long advocated the idea of combining biological and technological intelligence.

One of his visions involves introducing nanorobots into the body's capillaries – the smallest branches of our blood vessels – to connect the neocortex to the cloud.

Such a connection could theoretically expand the capabilities of the human brain and provide access to enormous quantities of information.

The neocortex is one of the most highly developed regions of the human brain.

Its functions include processing sensory information, language and aspects of movement and cognition.

It also plays an important role in conscious experience.

The combination of biological and non-biological thinking could fundamentally transform human culture and technology.

As Charles Darwin's theory of evolution illustrates, adaptation plays a crucial  role in the survival of species.

But if humanity itself assumes the role of a creator and develops an intelligence that surpasses mortal Homo sapiens sapiens, the consequences could profoundly influence the future of life on Earth.

The next stage of evolution has begun.

Anastasia Weimer