Robotics is an ever-evolving field that has the potential to fundamentally transform business operations at all levels, and the push is well and truly on to develop increasingly responsive machine intelligence.
While the boundaries are forever being pushed, the limits of robotic intelligence and where it will ultimately lead is the subject of wide-ranging speculation.
As it stands, the impact of new robotic technology is being seen across a range of sectors, from industry through to consumer applications.
As Tom Morrod, research and analysis executive director for the IHS Technology Group, outlined in a recent blog post, robotics in the past few years “has gone through astonishingly rapid development”.
Morrod says this is due to an overlap of a number of factors, including high-powered computing, compact and mobile components, along with big data machine-learning and low-cost 3D manufacturing.
Morrod points primarily to two aspects in drawing a line between physical devices of similar form to current robotics.
“The first is that unlike other comparable devices, robots have the ability to provide feedback to their human controller, using sensors to create haptic – ie touch – or remote input,” he writes.
“The second is that robots are able to process sensory data in order to make decisions on how to execute tasks, and in more advanced cases, be able to also define and elucidate the nature of the task to be performed.”
How do you define intelligence? While there are a number of methods designed to measure human intelligence, however in broaching the subject it is easy to fall into the abstract.
Morrod notes that in robotics intelligence is “conceptually complex”. Applying abstract definitions to machine codes presents the difficulty of being “too aware of the processes involved”.
“So it becomes very easy to understate or dismiss machine intelligence as distinct and distant from organic intelligence, simply because we already understand the details of the process,” he writes.
“If a human codes for the behaviour, it cannot be truly intelligent.”
However, Morrod observes self-learning structures, in which intelligence is not programmed but learnt, are challenging this perception.
“This process of generating intelligence is analogous to the way that organic intelligence develops through trial-and-error and then repetition, ultimately leading to innovative decisions — that is, decisions that are not predisposed,” he writes.
Machines are learning
‘Machine learning’ and ‘deep learning’ are two terms that are now being thrown around more often in relation to artificial intelligence.
According to Morrod, it could be that the “the crucial change in recent machine intelligence is the development of machine learning – more specifically deep learning”.
“The current stage for most artificial intelligence is pattern recognition from large volumes of text or visual data, such as the facial recognition algorithms on Facebook, the contextual recommendations of Google Assistant, or the natural language processing of speech by Amazon’s Alexa,” he writes.
“Recent machine learning systems use a process called deep learning, calling for algorithms to structure high-level abstractions in data via the processing of multiple layers of information. This occurs as machine learning tries to emulate the workings of a human brain.”
As with human intelligence, when it comes to AI it is difficult to pin down a definition. Moreover, as our collective expectations shift in concert with technology developments, definitions will likely also shift.
Morrod notes that recent achievements, such as the ability to recognise faces and objects, are “premised on a method that is still relatively explainable”.
“These achievements, while remarkable, also lack a higher order of intelligence demonstrable in abstract, non-material qualities like creativity, understanding or self-awareness,” Morrod says.
“In all likelihood, it is only a matter of time before these triumphs are considered mere computer know-how and not really AI. At that point, our expectations will then move even closer toward abstraction, which we currently find harder to define.”