From 600+ conversations with the world’s leading thinkers.
There's a compelling hypothesis in the field of machine learning called 'overfitting,' where systems become so tuned to recognizing patterns that they falter when faced with new, unexpected scenarios. Interestingly, some computer scientists are exploring ways to inject noise into computational models to keep them adaptable.
What we envision is humans using informal reasoning and intuition as a powerful guide, with formal systems then verifying those ideas. In this way, the formal system grounds high-level intuition.
We therefore see the drone exhibiting through software signs of the moral-affective function of 'guilt' when engaging in each mission.
Most people run around like biological robots, as if we are an algorithm not a being. We become the predictable outcomes of the conditioned reflexes of our nerves, constantly triggered by people in reaction to circumstances.
Because these systems don't see the world the way we do, they can extrapolate things in novel and unexpected ways that we haven't identified. Systems like Deep Mind's AlphaGo are not beating humans at games through speed and brute force, they're discovering new ways to play which we never conceived.
You can have the best AI in the world and the best robots in the world, but if they aren't integrated well with the humans, then you will lose.
We have a lot of phobias around algorithms. Sometimes this is justified, but in the main, it's like being afraid of cockroaches or spiders. Algorithms aren't spiders or cockroaches, they're an instrument and sometimes will outperform human judgement terrifically well – and sometimes won't. If lives are on the line and it turns out an algorithm reduces the noise of the human decision maker and the bias, then the moral case for using the algorithm starts to look really strong.
Unlike a machine, which processes myriad data points yet remains detached from meaning, we humans instantly ascribe significance to our sensory perceptions. The colour red transcends mere visual data, morphing into a spectrum of experiences.
I grew up in the age of space exploration, reading Asimov, Le Guin and Clarke, getting lost in the worlds they had created which were littered with aliens, robots and AI. For me, I think it was a combination of Star Trek, far too much Asimov and the niggling question of understanding ourselves better through computing technologies.
Currently, we find ourselves at a fascinating juncture technologically, where advancements have vastly expanded our capabilities to manage and interpret large-scale data. Computational methods have now surpassed human abilities in certain areas, a milestone that, until as recently as last year, seemed distant. This achievement indicates that the expense associated with intelligent data processing is decreasing.
These are technologies that are autonomous in many, many ways. They are independent in many, many ways – they have free will. They can replicate. And that makes a difference because then we teach them how to learn, but we have no idea what they will do with that ability to learn.
We're living in a world of increasing, exponentially growing computational power. Technology is always on, always available, and we're now moving into the quantum computing era – these exponential technologies are enabling artificial intelligence, robotics, 3D printing, synthetic biology, augmented reality, blockchain and allowing these technologies to converge, creating new business models.