Rainy street in Cambridge (Senate House Passage) with cyclist in the distance and sun making window patterns on wall

December 6, 2024, by Brigitte Nerlich

Making Science Public 2024: End of year round-up of blog posts

It’s that time of year again when I write my round-up of all the blog posts that I have written over the year. There were more posts than I expected. I have tried to group them into topics, some of which you might be interested in, others not.

This year, I mostly explored the ever-expanding landscape of ‘artificial intelligence’ and LLMs, but I did not neglect old topics either.

Climate change

The first blog post of the year dealt with the old topic of climate change, but even that post was AI informed. It relays a conversation with Claude (the Anthropic chatbot) in which Claude tried to tell me what I say about climate change and metaphors and becomes even more metaphorical than myself!

More serious discussions of climate change and extreme weather events, in this case the German floods of 2021, can be found in a post summarising an article that finally came out this year. It deals with the ways people in Germany tried to make sense of the catastrophe through metaphors and metonymies (mainly mud) and how that may have impacted attitudes to politics and, perhaps, action on climate change.

More serious still is the fact that there seems to be a movement to attribute extreme weather events not to ‘man-made’ climate change but to ‘man-made’ interventions into the weather in the form of ‘cloud seeding’. So I talked about that in a post entitled ‘Seeding clouds – seeding doubts’.

At the end of the year I wrote a review/overview of a new book edited by a collective of authors working at the University of Hamburg about plausible climate futures and how that plausibility is shrinking faster and faster.

Science communication and history of science

This year I didn’t write many posts about science communication but I penned a review of two fascinating books on the topic by Emma Bloomfield and one by S. L. Seethaler.

I also did a bit of historical digging into the life and work of the 19th-century polymath John Herschel and the history of photography in this post, which I have also turned into an AI generated podcast, which I love.

Genetics and epigenetics

Genetics, genomics and epigenetics are still on my horizon. So I wrote one post on machine metaphors in epi/genetics, especially train metaphors, featuring Julian Huxley on epigenetics, François Jacob on gene regulation, and Chris Curtis on gene drives.

And from there to epigenetics (the study of how cells control gene activity without changing the DNA sequence). That topic has fascinated me for a long time, especially the fact that so many social scientists and marketing people have jumped on the ‘transgenerational epigenetic inheritance’ bandwagon, a rather shaky one that breaks down when you look at it more closely.

Anyway, in this post, I report on an article I wrote with Alan Valdez from the Open University on the depiction of transgenerational epigenetic inheritance in the news and how images of happy families make it enticingly familiar. This is dangerous as the topic is still not really settled in science.

I wrote two posts about sickle cell anaemia. I used these posts to scope out a potential article on sickle cell anaemia, gene editing and identity, which, however, never materialised in the end.

I never forgot about metaphor, of course. In one post I discussed the pros and cons of the well-known blueprint metaphor in biology but also the less well-known one of postmen! I also explored again the long-standing issue of war metaphors used when talking about bacteria, dissecting in detail an article in New Scientist which was a real shoot-out.

Diseases and stuff

This  year, various diseases and potential epidemics/pandemics are on the horizon again, amongst them bird flu, which is now cow flu and has the potential to become human flu and Mpox, again, this time centred on Africa.

Inspired by work carried out by my colleague Pru Hobson-West, I wrote a post about vaccine hesitancy in Europe and the various words used for it. And in honour of a former colleague who worked on this topic, Nik Brown, I wrote a post on xenotransplantation, that is, the transplantation of organs, tissues, or cells from one species to another, especially to humans.

Organoid intelligence

I have been interested in the past in the visual and metaphorical framing of ‘organoids’ or organs like brains on a dish. This research is now fusing with AI and so I made a foray into something between biology and AI, namely organoid intelligence and hybrid robots, the convergence of living cells and tissues with artificial components. And so we come to AI.

Artificial Intelligence

Of course, when it comes to artificial intelligence one has to think about intelligence; and so I wrote one post about ‘intelligence’ and one about ‘superintelligence’ – from the divine to the digital.

I then turned the superintelligence post into a podcast using Google’s NotebookLM (Deep Dive).  That was a fascinating experiment. NotebookLM has by some been called “Google’s ChatGPT moment.”

Thinking about intelligence made me also reflect on how we humanise or anthropomorphise ‘artificial’ intelligence and, in the process, may be, dehumanise actual human intelligence.

My podcasts about superintelligence and the one about Herschel were not the only way I experimented with AI or rather LLMs/chatbots. I did it a lot and I then wrote a blog post about my experiences entitled ‘playing with Ai’.

I did more than playing though and dealt with a variety of more serious topics relating to AI. I wrote a post exploring the intersection of large language models and DNA models, as well as the interface between LLMS, meaning and maths. (Not that I understand any of it!)

I also explored two issues that loomed large over the past year or so in AI, namely AI safety and responsible AI.

And finally, I became fascinated by the magic of prompt engineering and linked that to the study of speech acts. That was fun. Less fun was thinking about AI and issues around truth, post-truth and post-fake….

AI and metaphors

Many of my forays into AI were, of course, metaphor related. I wrote one post entitled ‘hunting for AI metaphors’ ( that hunt is still going on) and another on AI, LLMS and an explosion of metaphors. That covered AI metaphors more generally.

I also talked with Claude about a particularly troublesome metaphor in biology, namely the machine metaphor and learned a lot from that ‘conversation’. In this context I also examined a metaphor introduced by Kevin Mitchell and Nick Cheney to replace older genetic metaphors like the blueprint metaphor, namely the genome as autoencoder metaphor – and again I discussed that with Claude who made some suggestions about science communication.

Prompted by a new book written Philip Ball on ‘How Life Works’, his rejection of old biological metaphors and his appeal for a new language for a new biology, I delved a bit deeper into the history of such an appeal. I went back to the 1960s when linguists, geneticists and anthropologists came together to explore language and biology and how they inspired each other. Perhaps it’s time to repeat that exercise.

Finally, I homed in (this time sans Claude) on a particular disturbing trend in metaphorization namely pollution, contamination and even collapse metaphors in AI, something that happens when chatbots feed on their own synthetic text. By the way, the word ‘enshittification’ also known as crapification and platform decay was selected as word of the year for 2024 by the Australian Macquarie Dictionary. It describes a pattern in which online products and services decline in quality and was made famous in an article by Cory Doctorow (see Wikipedia).

And back to climate change

The issue of climate change with which I began this year’s blogging doesn’t go away. I was just looking at the Valencia floods and the floods in Wales and many more when I saw that some colleagues tried to find out how chatbots portray the climate crisis.

That got me into conversations with various chat bots about the climate crisis. I was pleasantly surprised to find that they all conveyed the consensus science of climate change and its impacts.

The post reporting on this ends with a link to an AI generated podcast of my own conversation with Claude – all a bit meta and uncanny, but fascinating, especially as the bots encourage urgent climate action.

How long this will last, that is, how long the bots will stick to the consensus view on climate science and climate action is another question, given the many political changes happening all over the world which pose serious problems to how humans and bots may survive in the future.

Image: Senate House Passage in Cambridge, November sun making patterns of windows on walls of Gonville & Caius College

 

 

 

 

Posted in Uncategorized