Are we ‘summoning the demon’ with our embrace of AI? We asked Dr Steve Marsh, Founder and Chief Technology Officer, GeoSpock

Elon Musk famously declared we are ‘summoning the demon’ with the development of artificial intelligence (AI), and cited it as the most serious threat to the survival of the human race.

Coming from someone famous for cutting edge technological innovation, Musk’s disturbing predictions were taken to heart, and have contributed to an overall view of AI as a concept that should be feared. Add in sci-fi themes from the likes of Blade Runner and Dan Brown’s novel, Origin – where autonomous bots threaten the human race, and it’s easy to see where the dread of AI comes from. As AI capabilities progress, fears grow that fiction could soon become reality.

Real world developments have also fuelled anxiety about AI, from Facebook’s chatbots Bob and Alice creating their own language to negotiate with each other, to campaigns such as Stop Killer Robots, which aim to prevent the future development of fully autonomous weapons. On a more practical level, people are worried AI will take their jobs, with PwC research predicting that one in three (30%) of UK jobs could be at risk from automation within 15 years.

But can we find a way to harness the power of AI for good rather than evil? Musk clearly thinks so because since making his chilling remarks at a university symposium in 2014, he has worked to steer AI research towards positive outcomes. In 2015 he created OpenAI, a research company that aims to develop artificial general intelligence (AGI) in a way that benefits humanity and in 2016 he co-founded Neuralink, creating AI-based devices to be implanted in the human brain and ultimately help humans merge with software.

AI developments are largely constructive rather than destructive, with enormous potential for automating processes, managing big data, and aiding decision-making. PwC estimates productivity gains from AI will add $15.7 trillion to the global economy by 2030, and the British Government has so much faith in AI it is investing £75 million in 2018 alone. AI is already being used in many positive ways, such as AliveCor, which monitors heart rates and activity levels to detect inconsistencies and enable the user to take an EKG. And consumers are embracing AI in their everyday lives, with over 80% of TV shows watched on Netflix discovered via the platform’s AI driven recommendation engine.

In reality, while AI is continually developing, we are still a long way from the type of AGI where machines are truly able to think for themselves and potentially pose a serious threat to humanity. Today, even the most advanced AI can only perform the specific task it is trained to do and can’t apply intelligence more widely. As part of my PhD at Cambridge University, I built a custom super-computer for real-time simulation of human brain function, but even that is a far cry from actually replicating human intelligence.

The continued advance of AI requires extreme volumes of data, from multiple different sources, as well as the processing capacity to analyse data at scale. At GeoSpock, we are making significant progress in this very direction. When I founded the company in 2013 existing software was unable to keep pace with data generation speeds, whereas the GeoSpock platform is now able to query hundreds of billions of rows of data with sub-second response times. Our ability to conquer data beyond the trillion-point petabyte-scale facilitates huge advances in AI, allowing systems to be trained on the fly using extreme-scale data sets. Yet these systems are still guided and overseen by real people, and even without futuristic human-level AI, they present vast opportunities for humanity not threats.

Ultimately AI will continue to disrupt our lives, but we will see this happen largely in a positive way – used to drive smart cities, develop advancements in health services, and make great strides in other beneficial technologies. Elon Musk may still believe Artificial General Intelligence is ‘summoning the demon’ but, without extraordinary and as yet unforeseeable advances that allow the replication of real human intelligence, the demon simply won’t be woken – in the mean time I look forward to “simple” AIs drastically improving the human experience by reducing road traffic fatalities, shortening our travel times, and improving congestion and pollution in smart cities.

Leave a Reply