Machine Intelligence

For many centuries, humans have dreamt of intelligence in a form that is equivalent or surpassing to theirs. From the 16th century Jewish tale of the Golem of Prague, to the more modern day depictions of sentient robots mimicking humans in films. We have a fascination for what it means to have machine intelligence.

What many don’t realise however, is that with the advent of modern mathematics, neuroscience and computer science, machine intelligence is already here!

In fact, it has been here since 1956 when the term “Artificial Intelligence” was first coined, and the most fundamental assertion of the field was established.

"Every aspect of learning or any other feature of intelligence can be so precisely described that a machine can be made to simulate it"

That assertion was made at the 1956 Dartmouth Conference, and since then the field of machine intelligence has been growing exponentially. However as with any exponential growth, the initial period has been long and slow, with little visible progress and results. But as the human race rounded the corner of the 21st century, so the power of computers has dramatically increased. This increase in computing power has allowed us to take the theorised mathematics of machine intelligence, devised in the 1970s, 80s and 90s, and apply them on vast scales. We have rounded the inflection point on our exponential curve, and machine intelligence has taken off.

What does this translate to? Currently the average human has numerous interactions with machine intelligence every day. With each interaction these systems are getting smarter and smarter. Every time you shop online, every time you send a message or email, and even walking down the street, multiple systems are learning, analysing, and evaluating the behaviour from every interaction and every move.

These machines have not been programmed in the traditional way of fixed rules and fixed behaviours. They use large systems of neural networks to make their own rules, in essence programming themselves to learn the most optimal method to complete a task. Just like a human brain does. This is because these networks are, in fact, based on the intrinsically complex workings of biological brains. The difference is biological brains are constrained to working at biological speeds, constrained to evolving through natural evolution and constrained by the scale they can take up in a physical body.

Machine intelligence is unconstrained by these bounds. Neural connections in an artificial brain are transmitted at the speed of light. Individual electronic neuron processes can be infinitely faster than biological ones. And the scale of silicone chips connections is not limited to fit in a human skull, but instead can span vast acres of data centres.

How does this tie in with the future and the present? We are currently passing through the last part of the inflection point the exponential growth curve of machine intelligence. At the same time we are producing the largest volume of data that has ever been produced in the whole of human history up until this point. Every aspect of our lives are covered. Within that mass of data there is meaning. Incredible insights can be derived by combing the huge, infinite dimensions from all of this data.

In the past, skilled human analysts worked at analysing data in this very human way to derive insights from it. But their capacity was limited. They could usually only analyse a handful of dimensions of data at a time before their methods became complicated and cumbersome.

This was fine at the time, because the data itself was limited. This is no longer the case. With the vast amount of data now being collected and available, the number of dimensions that can go into deriving useful insights for something; be that a medical diagnosis, a stock market event, or even predicting the next thing you might do, are unbounded.

This is where the machine intelligence revolution has been happening, is currently happening, and will only continue to grow. Machine intelligence is bringing unparalleled efficiency gains. And whilst we have a way to go until generalised machine intelligence is created that can truly have a human-like consciousness, narrow-domain machine intelligence tools are already being used to augment human decision making. Effectively turning humans into super-humans!

The future of humanity is changing. And whilst nobody can yet predict what will happen in the next 20-50 years we know that changes will have to be made to accommodate this new intelligence.

Listen to our Founder and CEO giving his thoughts on the future of artificial intelligence: