Some brief thoughts on scholarly publishing and climate change.

Categories: climate crisis ai

Who’s job is it to address the crisis of climate change? This question came up at an STM Tech Leaders workshop that I attended earlier this week? The answer is that we all are, and in that context we had a very interesting conversation on the topic.

This is the first time that I’ve seen this topic come up in a forum like this, so a bit thank you to Dave Smith from the IET for pushing the topic forward.

Oh man, this is such a big topic, but we have to start somewhere, right?

We had a short, and very useful conversation. I hope there is more space made for this in the future. There has to be. At the recent SpontOn London looking at publishing and the UN sustainable development goals one of the speakers outlined how as a species we have destroyed every habitat that we have colonised, and that the most appropriate description of a species that does that is a parasite. She also pointed out that all of our science and technology has been a force multiplier for that parasitism. (She made a call for us to focus on embedding sustainable values into our education, at that very moment where we equip ourselves with all of this knowledge we have generated). Well I digress, but it’s clear that this work we are engaged in - the knowledge creation work - has a pivotal role to play to address climate change. Furthermore we are working in companies of not insignificant scale and so we have an opportunity to embody best practice in a meaningful way to help make a difference.

SpringerNature told us about how they had made the biggest impact on their carbon footprint last year by re-configuring their Indian offices to be more energy efficient.

Ours is an industry where flying happens a lot. How might we reduce this need? How can we not only cut down on travel for internal staff, but also help the editorial boards we work with reduce the need to fly? The clear answer here is just to fly less. I’m not sure that AR/VR are going to be a huge disruptor, but they could be, they could be. It’s not a crazy thing for us to keep experimenting with these things.

Paper remains a bit factor in our footprint. Let’s go online only already. (I just don’t understand the print side of our industry, and I keep forgetting about it, but there are a lot of atoms and not just bits involved here. Perhaps we could start to tell stakeholders about the carbon price of their print editions, perhaps thin information would be another nudge to get them to go e-only?

We do a lot of compute. We are all excited about AI. AI is compute intensive. It’s probably not a significant % of the kind of compute we do as an industry, which is probably dominated by maintaining databases, and by production workflows, but AI is compute intensive. Are there ways to move to shared models that could be built that cover the broad use-cases that are of interest to us, and that could make a big dent in the cost of building the same, or similar, models over and over again?

Of course, we help to publish the damn research. What could we do to help make this key work more sticky for consumption by the general public? We had a great chat about whether AI could help with summarisation in a way that easy to digest versions of articles could be created in a cost effective way? Say for example by training a language model on simplified-english wikipedia, and passing abstracts or full texts through this model? That’s one approach to the idea, with a bit of thought I am certain we could come up with others.

At least for now though, I think we should all take this away to our own organisations. Is this being debated internally? Are there action groups? Is it making its way into strategic planning decisions? How can we make that happen? Who should we talk to in our organisations? I’m going to start by forwarding this article to some of the planning groups I’m involved in within SAGE. We know and accept that climate change is a social justice issue, and we know that lowering our carbon footprint has a benefit to our bottom line, we just, I think, need to keep pushing on this to allow ourselves to get continuously better at addressing this issue.