Predicting the future’s hard but there are two global trends that seem fairly certain. Digitalisation and climate change are likely to shape our future more than anything else that we can see at present (short of catastrophe). How are they linked? Or, to put it another way, why aren’t they linked more than they are?
This week, the problem. Next week, a suggested approach.
An existential threat
When I was young, the existential threat we feared – the thing that could destroy humanity, or at least the world we knew – was nuclear war. That threat’s still there, though people have acclimatised to it.
Today’s most feared existential threat is climate change: longer-term, insidious, with processes that are hard to fathom. Not surprisingly, it’s the young who’ll live with it that care about it most.
Current trends aren’t optimistic: carbon and other climate-harmful emissions continue to grow. (In what follows, I'll use "climate emissions" for convenience.) Most governments agree that they should fall but political priorities lie with short-term economic growth more than long-term sustainability.
Which is a problem for ‘sustainable development’ and its UN-sponsored Goals. Their aim’s to achieve balanced improvement in economic prosperity, social welfare and environmental stability. But their target date is 2030, before the impact of climate change is going to be really felt. Not surprisingly, therefore, the environmental side of them gets less attention than it needs.
What has this to do with digitalisation?
Not everything, by any means. What happens with climate change will be determined by political will, by demographic pressures and by the extent to which international cooperation overcomes the vested interests of individual governments, businesses and communities.
But quite a lot, nevertheless. Digitalisation will affect the way most things are done between now and 2030, even more twixt now and the middle of the century when climate change really begins to bite.
These two trends – digitalisation and the growing climate crisis – tend to be considered separately, but they will interact, and do so more and more as years go by. Global policies for the future need to understand that interaction and, as elsewhere, seek to maximise the gains and mitigate the risks. At present, I’d say, they’re not considered jointly near enough.
What the industry says will happen
The ICT industry likes to think it has solutions, but all technologies have negative as well as positive impacts.
It’s argued three things, mostly, as solutions:
that ICTs improve efficiency, so that less energy consumption’s needed to achieve the same result – in manufacturing processes, in managing lighting and heating within buildings, in smart systems that manage traffic flows, etc.;
that they improve monitoring of potential problems, allowing them to be addressed more quickly and effectively, again reducing costs including carbon costs;
that they allow different ways of doing things, such as videoconferencing and telecommuting which, they say, will reduce carbon too.
As usual there are ifs and buts here. I’ll come back to what’s really happening in a moment.
Assessing carbon impact
I’ve written previously about different strands in the relationship between ICTs and carbon:
first order or direct effects – the carbon cost of making and using ICT equipment (data centres, computers, phones, etc.);
second order or indirect effects – the carbon cost or benefit of using them to do things which we already do or which we couldn’t previously do but can today;
third order or societal effects – the carbon cost or benefit of changes in society – changes in trading patterns, for example, or in the ways that people work and play, or human settlements – that result from pervasive future use of digital technologies.
The evidence from this is complex. First order effects obviously add to carbon costs. The industry believes that second order effects are beneficial, though the evidence is much less clear (see below). Third order effects are far harder to predict.
At present the ICT sector accounts for rather more than 2% of global carbon emissions – a little more than aviation – and that figure’s rising faster than any other sector. The rate of growth is also likely to increase as digitalisation spreads (see below).
What’s really happening
What’s really happening’s more complex than the industry admits.
First, we haven’t seen those carbon cuts from different employment patterns. Videoconferencing and telecommuting have grown but haven’t taken off as much as had been hoped. Instead, we’ve seen the growth of gig platforms that individualise service provision, increasing carbon use. Air travel has continued to grow as business flyers prefer meetings and prosperity increases tourism.
Second, the most important effects on emissions may be rebound effects. Greater efficiency in energy consumption doesn’t necessarily lead to less consumption; often it leads to more.
Efficiency reduces costs to consumers, leading to more use of relevant devices. Think of how many ICT devices you now use compared with those you used ten years ago. Think of how often you leave them powered up rather than powered down.
More efficient traffic management may well increase rather than reducing use of private vehicles – because they become more, not less, convenient. Less time at work means more time at home, which is as likely to increase energy consumption per head as to reduce it.
This is not an argument against digitalisation, but the case for considering its impact more holistically and identifying policies to maximise advantages and mitigate disbenefits. Measuring what’s actually happening is crucial here, and far more complicated than the simple balance between first order costs and benefits of improved efficiency that the industry’s tended to emphasise.
And what’s going to happen
The need to measure this will be even more important in the future because of the enormous increase in digitalisation that’s now underway. More people using more devices for more of the time, generating more data that requires and gets more (and more complex) analysis in more data centres. More data traffic, and yet more data traffic, using more connectivity, constantly.
This digitalisation of everything won’t be carbon neutral. Global mobile data are currently estimated to be growing at a compound rate just under 50% a year – from 11.5 exabytes per month in 2017 to 40.5 this year and 77.5 in 2022. Every byte of data needs an extra bit of energy.
Big data, the Internet of Things, artificial intelligence and machine learning will require much more. Think, for example, of how much data will be generated, analysed and communicated in a transport system that’s based on driverless technology.
That 2% plus of carbon emissions for which the sector’s now responsible’s expected to grow to 4% and more within the next decade, without (for reasons above) a compensating fall resulting from it elsewhere.
Addressing environmental impact
Lack of attention to predictable outcomes such as this imperil the future – for the ICT sector itself as well as the environment. We’d be in a better place today if the pioneers of the internet had paid more attention to cybersecurity; a better place in terms of privacy if we’d put rules in place to control the exploitation of personal data by data corporations before they’d built their business models round it.
Information technologists have played far too little attention to environmental impacts when designing IT systems. Bitcoin illustrates the issue. The world’s first major cybercurrency uses a technology whose energy requirements are wholly unsustainable at scale. They already match those of a sizable developing country and are growing fast, yet few people use it.
I am not making the case here for less innovation, but for innovation to be more responsible: for more thought to be given to environmental impacts (not just on carbon, but also on natural resources, on redundancy, on waste) in the design and development of a technology that’s meant to help not harm tomorrow.
Next week I’ll look at what more responsible innovation might look like in practice.