Let´s talk about AI

Big investments are being made in Artificial Intelligence (AI) right now. What society can and would like to use AI for is less frequently discussed. A more detailed dialogue is needed.

Artificial Intelligence is described in various contexts as a key technology that is going to change society and industry in a fundamental way. Different research funding bodies are providing big sums. The goals of these investments are said to include to benefit Swedish industry and to fuel industrial development.

I think AI offers enormous potential. For this very reason, those of us within research also need to discuss what society can and would like to use AI for.
Does it matter which industry will benefit or are there certain applications that are of more or less interest?
Are there risks with AI we need to become aware of, to further research and discuss?
How can AI contribute to sustainable development and is there a risk that it will prevent this? If we are to address these and other questions, research into AI cannot simply be the domain of the most closely related scientific subject areas, it must also embrace other relevant areas.

In association with international discussions on the opportunities and risks of AI, researchers and other stakeholders have formulated a number of principles for AI research. One of these principles concerns the goals of the research: “The goal of research into AI should not be to create random intelligence but benign intelligence”. Another principle states that “Investments in AI should be accompanied by the financing of research to guarantee the benign use of AI. Research should include thorny issues within computer science, economics, politics, law, ethics and social science…”. These principles could be interesting starting points for further debate.

If you compare the principles quoted above with the goals of the substantial investments within AI that are being made today, there appears to be daylight between them. It is therefore important that there is increasing debate both within universities and beyond, concerning what AI research is needed, how it should be financed and what the purpose of the AI research should be.

Tip of the week:Follow the webinar on how CO2 emissions can be halved by 2030, from 13.00-15.00 on 11 January. Researchers from KTH and other universities present The Exponential Climate Action Roadmap. For more information and to register, click here.

How aviation can reduce its climate impact

Air travel accounts for an increasing share of the climate impact of Swedes. Flights are currently responsible for just over 10 percent of the carbon footprint attributable to Swedish consumers. That is around about the same as private car use in Sweden.

To reduce the climate impact of aviation, several different types of measures are required:
• More fuel efficient aircraft.
• Fuel with a lower climate impact, such as biofuel or hydrogen. However, these fuels will not totally eliminate climate change effects as they still give rise to so-called high-altitude effects and emissions during the production of these fuels.
• Higher occupancy rates in aircraft.
• Better organisation of air traffic (e.g. greener landings, shorter flightpaths and flightpaths that reduce the high-altitude effects)
• Reduced air travel.

Actions within all these areas are needed to reduce the climate impact of aviation. The first four points show the effects on emissions per passenger kilometre. There has been substantial development over the past few decades that has helped to significantly reduce emissions per passenger kilometre. It is important that development work continues, and these areas are likely to offer big potential. However, the number of flights is increasing at the same time, which means total emissions have increased. If we are to be able to reduce total emissions, the number of flights must therefore also be reduced.

This week I took part in a seminar arranged by the Expert Group on Public Economics (ESO), on climate policy that inter alia, addressed the climate impact of aviation. Politicians from the Swedish Social Democrat, Conservative and Center parties took part in the panel discussion and they were all agreed that the price of air travel will increase moving forward, via requirements for a mix of biofuel and/or taxation. It is therefore of interest that the final point, reduced air travel, will also be affected. However, the question is whether the above will have a significant enough impact on emissions, or if additional instruments will be required.

Tip of the week: Go to a lecture on batteries or TEDxWomen Conference a lecture Where will Future Paths lead us? All at KTH on 6 December.

How can we deal with unknown risks?

We cannot always foresee the consequences of new technology. What kind of rational and intelligent decision-making do we need in such situations?

I was reminded of this question during the KTH Sustainability Research Day last week, when we learnt about a number of case studies of what kind of penetration and impact research at KTH has had or can possibly have on social development. The first example was the fridge. That was a product that has had an enormous impact on our everyday lives and on the world’s food supply.

Fridges were developed in part by KTH students in the early 20th Century. In the 1920s and 30s, freons started being used as coolant media. They were considered to be a fantastic product for many years. They were stable, non-flammable, and moderately toxic. However, the first suspicions began to arise in the 1970s as to whether they were so stable that they could spread to the stratosphere, that is to say the layer tens of kilometres up in the atmosphere. When freons reach that level, they can break down ozone which, in turn, is needed to protect the earth against harmful ultraviolet radiation. Some ten years later, these suspicions were confirmed by the discovery of the so-called ozone hole above the Antarctic.

A reduced level of ozone in the stratosphere can result in serious health effects, such as skin cancer, and also affect the eco system. Demands for a ban on freons came pretty quickly and this time, within a few years, decision-makers around the world managed to agree on the so-called Montreal Protocol in 1987 to phase out freons. That was a wise decision. Large parts of industry protested, but alternatives were able to be quickly developed and most of us can go into the kitchen today and see a fridge that performs well even though we do not know which refrigerant it contains.

Many aspects of the history of freons and the ozone hole provide lessons we can learn from. One thing we can ponder about is whether we could have foreseen that freons could have the effect they have on the ozone layer. Could we have understood what would happen to the freons when we scrap a fridge, and did we know enough about atmospheric chemistry and the special conditions that apply above the Antarctic at certain times of the year? Many people would say no, we could not have foreseen that. And the same also applies to many other classic environment problems such as DDT and PCB for example. It was very difficult to predict the very specific spreading pathways and effect mechanisms these substances have.

But if we cannot foresee the consequences of a new technology, how should we then deal with its risks when the decision is made? One lesson is perhaps to watch out for warning signs.

When research signals that there can be a problem, these signals need to be picked up and investigated properly. In such situations, there are special interests that want to tone down the risks, which makes it even more important to investigate them in a neutral way. Another lesson is the need to find opportunities to make decisions quickly enough. In the case of freon, it did not take that many years between discovering the ozone hole and phasing out fridges with freons. These decisions faced opposition but were implemented anyway and that was important. A third lesson is to be especially vigilant about substances that are persistent, that is to say, difficult to break down. Freons, DDT, PCB, plastics in the ocean and carbon dioxide are all examples of substances that are persistent. Which means they can spread far and wide and appear in places that are unexpected and continue to have an effect long after they were used. It then also becomes difficult to deal with the problem as it takes a long time for these substances to disappear.

These lessons come from the environment area above all. However, I believe we have reason to think long and hard about how we can deal with unknown risks in other areas, too.

Tip of the week: The CrosscutsFilm festivalin Stockholm 23-25 November organised by the KTH Environmental Humanities Laboratory (EHL) in Stockholm

How Swedish consumption impacts the international environment

Consumption in Sweden affects the environment both here and in numerous other countries around the world. However, it is now possible to estimate this environment impact and take steps towards more sustainable consumption and production.

Achieving sustainable consumption and production patterns is one of the global sustainability goals. We also now have new methods to estimate the environment impact of Swedish consumption. These have been developed in cooperation between researchers at SCB, KTH, Stockholm Environment Institute, Chalmers University of Technology and the universities of Trondheim and Leiden. Previous analyses have primarily focused on climate impact, but methods are now also available for other air pollutants, the use of hazardous chemical products, water, land and materials.

In the majority of these aspects, the environment impact of Swedish consumption is greater beyond Sweden’s borders than within the country. For example, 65 percent of greenhouse gas emissions from Swedish consumption occur in other countries.

Around 80 percent of hazardous chemical products usage linked to Swedish consumption occurs abroad. This also covers pesticides and antibiotics used in veterinary medicine that are far more prevalent outside Sweden. Emissions of a number of hazardous substances also arise to a greater extent in other countries. Important countries with regard to the use of pesticides include The Netherlands and Brazil, Germany and Spain for veterinary medicines and China and Russia for emissions.

There is a long list of possible measures available to reduce Swedish consumption related emissions. One such area that is also relevant to KTH as a public body is to use procurement as a tool to specify stricter requirements. Another is to reduce the use of fossil fuels as this will not only lead to lower emissions of greenhouse gases but also of a long list of other substances. The choice of building materials and food products also plays a role.

Tip of the week: Take the boat. I was in Helsinki the other week and went there and back by boat. It worked really well.

 

Something that is socio-economically profitable can be unsustainable

Sometimes, socio-economic profitability and sustainable development are at odds with each other. One interesting question then is what should take priority?

To be able to compare costs and benefits that arise in the future with those arising today, some form of discounting is often used, interest, in other words. This enables you to calculate future costs and benefits at today’s value.

Naturally, one important question then is what discounting rate should you choose. Do effects in the future have a greater or lesser value than those happening today? If you consider that future effects are less important, you choose a high positive interest rate, if you feel they are more important, you choose a negative interest rate.

In current socio-economic calculations, a rate of 3.5 – 4 percent is often chosen. This means that the value of future effects falls rapidly over time. In 2118, 100 kronor would be worth the equivalent of 1.60 kronor today. In a report from the Swedish Scientific Council for Sustainable Development to the Swedish government, we write that a high discounting rate is incompatible with long-term sustainability if it leads to measures that are necessary for sustainable development being rejected because they do not compute as being profitable from a socio-economic perspective.

That socio-economic profitability and sustainable development can conflict with each other is an important insight. The question can then be what should take priority. As a public organisation (such as a university) you can fall back on the form of government (one of Sweden’s constitutional laws) that states that the public sector should promote sustainable development. We also write in the report from the Scientific Council that a rate that is low enough to be compatible with long-term sustainability ought to be chosen.

Tip of the week: DN debate about the role of universities in a climate change reset.