The shifting sands of tech
This post is about the shifting sands of technology we find ourselves in as simple flesh and bone humans trying to deal with the landscapes of media and information in our datafied lives. Nothing is certain, nothing is permanent. It is actually a bit terrifying.
First Wave
Web 2, at first seen as embracing the democratisation of knowledge and communication, soon collapsed into private walled gardens and has become known as the Surveillance Internet model. It was the realisation that someone had to pay for the infrastructure, someone had to have ideas about how to manage the explosion of human voices suddenly on the web airwaves, all clamouring for attention. And some way of making money, the monetisation of digital democracy. Voices appeared on the fringe in academic debate, people like van Dijck's Culture of Connectivity in 2013, more recently Platform Society from van Dijck, Poell & de Vaal, 2018, and Srnicek's 2016 Platform Capitalism, warning us about what had happened, and that it was a very bad thing indeed.
Second Wave
As we can see approaching fast towards us, things have occurred that will change the landscape forever. Public access to Artificial Intelligence chatbots (aka turbo confabulation), Artificial 'art' that includes exact replica photography and high quality generated illustration, and then, the Twitter collapse. All these things are far more significant than most people might believe. Cosy homestead opinions about these issues are largely honed from the semi-nice world of pre and post Web 2 ideas but that world was a long time ago. Remember those images from 'e-learning' academic texts that show all that really old technology? Most of it is in the dustbin of history, a fleeting blink of irrelevance devoured by platform capitalism. Remember Rita Kop in 2012? She quoted this: "The universal index is the shared heritage of humanity. It ought to be owned by us all. No corporation or nation has the right to privatize the index, commercialize the index, censor what they do not like or auction search ranking to the highest bidder. We have public libraries. We need a public search engine" citing White, 2010. She was right then, but no one was listening. It is surely a curious thing to watch experienced academics rush towards an unknown techno-fantasy future that they believe they must adopt and support, when in fact they have little real comprehension of what they are buying into, and it's contentious to even suggest this. But there's a terrible simplification of the landscape and a total absence of criticality. Remember also that many of the points used to sell us these technology adoptions promote concepts sold as 'new' that really aren't - that personalised learning, dialogic learning or knowledge rich learning content have never been done before and herald a fantastic new efficient learning future. Frankly, it's rather insulting.
The pervasive deepfake
The dangers of the implications of artificial intelligence large language model interpretations and artificial intelligence image generation are complex and real, they are not a conspiracy of naysayers who wish to return to a luddism of denial. You cannot unknow what you know, as the creators of Dolly the Sheep probably once said. But as we rush to thinking that using unreal war photography of soldiers in the trenches is 'useful', or to use fake nature photography 'makes no difference and saves time', or that chatbot regurgitation from a confabulated mishmash of unknown sources means 'more efficient learning', have we lost our minds? The threat of pervasive deepfake is now upon us. Most ordinary people don't think about whether a photograph of a bird in a bush is real or not, they assume it obviously is. But it won't be. When we see photographs of soldiers in the WW1 trenches in a textbook or a webpage, we assume they are real, but they won't be. When many graphic illustrators who have previously enjoyed work from books, magazines and websites are all made redundant, we will just assume they need to go with the times and get a new job, that it doesn't matter because it doesn't affect 'us'. But it will affect you. It will diminish every aspect of your own perceived reality, because your reality will have become consumed by deepfake. This is not really a very exaggerated postulation.
Possible new futures
While the feeling of uncontrolled megalomaniac owned AI domination will throw us all into a sci-fi dystopian deadzone remains 'possible', there may be light ahead. Open source Chat models have given us some hope. Hello Dolly, recently shared by Databricks, is one such thing. Their strapline is 'Democratizing the magic of ChatGPT with open models...', and it cost $30 to set up (AFAIK), using an older LLM. Results were comparable with the billion dollar counterparts. Just by existing, it proves you do not need millions of dollars to implement this kind of tech. Maybe public (tertiary?)education, as one example, should be investing in its own 'instance' of chatGPT, moderated by the academic community. It seems a simple idea, but no one has even mentioned this kind of alternative. As to AI Art, the issues are different, yet the same. A major challenge of all AI/LLM technology, is the rights to the intellectual property being scraped. As yet we are probably at the beginning of solving that, and its not much of a protection to just use robots.txt.
It is foreseeable that there is an easy escape in terms of social content and communication. The Fediverse offers a fresh vision of thinking about the globalised Internet of people, knowledge and communication. It is not a new idea but has gained traction since the Twitter meltdown currently happening, instigated by Musk's takeover and the de-public investment status. The private fiefdom of a globalised communication platform is a very dangerous precedent, and daily news about what is going on signals a new phenomenon, where individual communication channels only exist at the benign edict of the dictator. The sudden API functionality withdrawal, verification of accounts at cost, suspension of countless accounts, openly slandering vulnerable minority groups, permitting misinformation and propaganda about the Russian invasion of Ukraine, monetising the control of flow of information in favour of those who pay for higher profiling, labelling NGO and public bodies as 'state-affiliated' and then 'government funded', arbitrary icon changing, endless tinkering and foolery. An insane set of actions for any serious company to be undertaking, let alone an important global communication channel. There is as yet, even after more than two decades of digital life, no legal mechanism to control this random manipulation of global communication. It is a travesty, but it is what it is. So, the Fediverse beckons.
The Fediverse offers decentralised servers and networks that are interoperable, thereby automatically safeguarding the ownership of knowledge at grand scale. In layman's terms it means no one can own all of it, it stops being a walled garden, and individuals can move their followers and following to any new server home. Civic and private ownership can co-exist in this eco-sphere, and co-ops can own the servers that their information is hosted on, yet have a fully open door into everyone else's servers. It is a much more equitable landscape, a global network of interoperable information. Legal issues regarding data and geographical terrain can be cleanly administered too, which I think is one of the reasons why Meta have made announcements about setting up their own federated network. Zuck has probably realised that Facebook itself must decentralise in order to survive the post GDPR world and this is a very good thing, in my view. There will always be privately owned servers, and whether or not they choose to be interoperable with other federated platforms remains to be seen, but for now, this is the only way forward.
References
Kop, R. (2012). The Unexpected Connection: Serendipity and Human Mediation in Networked Learning
Full size image (click)
A generated image of The All Seeing Flame, made with simple generative art Windows desktop tools
