Set a Terminator to Catch a Terminator?

With intellectual heavyweights such as Stephen Hawking weighing in on the debate about whether the recent explosion in artificial intelligence (AI) will send us hurtling towards a post-human future, it is unsurprising that there is an air of concern over whether the rapid development in robotic technology holds hidden threats for the human race. While we have been able to buy pocket calculators for just a few pounds which can surpass our mental abilities since the seventies, the last thirty years have witnessed an immense acceleration in AI development, producing machines with the ability to learn, communicate and interact with us.?

Yet the field of robotics is still currently at a similar stage to computers in the sixties, with unbounded potential to develop in a very short amount of time. While computers have the ability to trounce the world’s best chess player and we have grown to utterly rely on them for countless menial tasks, there is still no robot with the dexterity of a five-year-old child, and their ability to sense their environment remains limited. Despite this, there are still many people who fear we are on the brink of, or arguably already in the middle of, ‘singularity’; a term coined by mathematician and author Vernor Vinge to describe the inflection point when machines outsmart humans, possibly resulting in “the physical extinction of the human race.”

Vinge’s 1993 essay titled ‘The Singularity’, argued that “we are on the edge of change comparable to the rise of human life on earth.” According to scientists, humanity’s future could be similar to that of the mitochondria; once independent organisms which were absorbed and incorporated by an ancestral cell called Lokiarchaeote. Over evolutionary history, the mitochondria allowed cells to take over all functions that they used to perform until they only produced energy. It has been suggested that as humans continue to transfer more and more tasks to machines, including examples considered to be fundamental to humanity such as caring for the elderly, that we are becoming the mitochondria of our machines. All machines will require from us is to initially provide the energy to switch them on. It has been predicted that where mitochondria-hosting cells instigated an extraordinary expanse of multi-cellular life forms which had never been possible before, human-hosting machines could allow humans to transcend their natural limitations and drastically alter the post-human being. Whilst this is a fascinating prospect, it creates the possibility of making human input redundant and as a result, evolve ourselves out of existence.?

This fear of technology is nothing new. We have seen it played out in all forms of literature and film since Mary Shelley’s ‘Frankenstein’ in 1818. But why are machines something to be feared and not celebrated? A much more optimistic view would be similar to that of Kevin Kelly’s in his article ‘The three breakthroughs that have finally unleashed AI on the world’, which focusses on all the benefits that cognitive computing systems can have in the future. Kelly states that AI systems are tools that enhance human capabilities and can be used to drastically improve current conditions. Similarly to electricity, cars, aeroplanes and computers, AI will “enliven inert objects” and that “everything that we formerly electrified we will now cognitize”, ultimately enhancing us individually as people and collectively as a species.?

Humans are prone to exaggerating probabilistic dangers as a form of self-defence, developed during a period of human evolution, reflected by the negative headlines that dominate our newspapers. We can overestimate the likelihood that we will have computers smarter than human beings and exaggerate the danger that these might pose to the human race, but in reality the development of intelligent machines is likely to be a slower and more gradual process than sudden extermination by cyborgs. While it is sensible to remain cautious of potential threats, I think that if our main fear is of machines with superhuman intelligence, then an apparent solution is more machines. Laws to control the development of AI have been suggested, but the possible advances are so compelling that it is unlikely they will be passed or adhered to. If humans are capable of engineering technology with AI, what is to stop them engineering ways to prevent consciousness in them, or machines that regulate development? This idea forms the concept of setting a terminator to catch a terminator.

In the financial markets computer-trading platforms are already placing trades faster than human regulators can track them, let alone regulate them, which leads to the so-called ‘flash crash’ effect. This is a huge threat to the foundations of our financial systems and will necessarily give rise to a solution that can only come from the machines themselves, as the only way to rein in the computers is with other computers. Regulation technology in financial institutions will reduce costs of meeting compliance requirements, de-risk operations and remove many of the headaches related to staying current.?

In a world where we are moving towards machines performing faster and more complex financial transactions on our behalf, we are forced to turn to those very same machines in order to police them, and the irony of this is not lost. Whilst artificial intelligence undoubtedly sparks a certain level of moral discussion and will prove to be disruptive in the future as it makes major modifications to the employment landscape, the benefits of AI will hugely outweigh the disadvantages and allow humans to achieve feats in the next few years that are unimaginable even now; it is an extremely exciting era. In fact, it is highly likely that we will wonder how we ever lived without it.?

Laura Elliott, October 2015