Science fiction writers have had no end of fun imagining alternative realities and parallel worlds. Today, we are building them in digital form.
With the number of connected devices forecast to grow to 42bn by 2025, according to research group IDC, we are rapidly entering the era of “hyper-data”. Each of those devices emits a constant stream of data, enabling us to build a digital cloud that will metaphorically encircle our planet. We can, to use the jargon, create “digital twins” of the real world.
Last week in New York, Microsoft’s chief executive, Satya Nadella, described the benefits of those digital twins, which he labelled one of the biggest trends in technology. Referencing Robert Gordon, the economic historian who has highlighted the stalling of innovation, Mr Nadella acknowledged that many economies were experiencing sluggish productivity growth. But Mr Nadella argued that increasingly sophisticated digital representations of the real world, enhanced by mass computing power and machine learning, would optimise information flows and catalyse economic growth.
Digital twins could also yield significant environmental benefits. “Bits will help atoms be more efficient by bridging offline and online,” he said.
To take one example, he pointed to UK retailer Marks and Spencer, which is increasingly using in-store sensors to create digital twins of its retail space. It can then use those data models to optimise its stores’ physical layout, monitor the temperature of its meat coolers and keep an eye on queues at its checkouts.
Similarly, factories can run far more efficiently and reduce waste if their operators are able to twiddle knobs in digital simulators and test ways to improve the run rate. “That changes the productivity of the manufacturing plant,” Mr Nadella said.
Others talk about digital twins representing the “third wave” of the internet, after search and social media. The online world will evolve from a consumer internet to an enterprise internet, profoundly affecting the way we run roads, railways, airports, mines, factories, hospitals, power plants, distribution centres, construction projects and cities.
“The real world has always been offline. Now we are putting it online,” says James Dean, co-founder of SenSat, a London-based geospatial data company backed by China’s Tencent that is using drone and satellite imagery to map infrastructure. “If we can create a digital representation of the world and we can add some funky analysis then it gives us digital superpowers.”
Mr Dean argues that digital twins can transform the construction industry, which has an appalling productivity record. SenSat is already modelling the route of the proposed HS2 railway line running from London to the north-west of Britain, creating some 18bn data points. That should help builders monitor the controversial project’s progress to ensure smoother workflows, vital considering the project’s big and growing cost overruns.
The vision behind digital twins is inspiring, but the unsettling experience of the consumer internet over the past two decades should alert us to the potential dangers.
Asu Ozdaglar, a professor at the Massachusetts Institute of Technology and deputy dean of its new Schwarzman College of Computing, says that all such complex systems are inherently unpredictable. “The excitement around how these technologies can transform society is very real,” she says. “But we should also be aware of their possible negative consequences.”
The obvious concerns are security, privacy, surveillance and ethics that need to be addressed before these systems are deployed. She warns that “cascades” of problems can also result from extreme interconnectedness.
Take car navigation systems, such as Waze, for instance. Although they promise to reduce delays for drivers, they sometimes route traffic through residential streets where children play. They can also prove counterproductive by encouraging herd behaviour. Drivers can flood “quicker” routes, worsening congestion. Sometimes less information is more useful.
“You have to build carefully designed models and look at the empirical evidence. Whenever you provide incentives, you get some unintended results,” Prof Ozdaglar says.
The explicit purpose behind MIT’s Schwarzman College is to blend the insights of computer scientists and social scientists to address such complex challenges. Its aim is to produce “bilingual” experts.
Industry must do the same. The consumer internet, built on the mantra of “move fast and break things”, has smashed too much crockery. The enterprise internet must continuously prove its social value in the real world.
Follow John Thornhill with myFT
Letter in response to this column:
Get alerts on Technology sector when a new story is published