I’m currently working on the End of Module Assessment (EMA) for my MSc in Systems Thinking which involves pulling together a lot of learning over the past few months. I’ve captured a lot of it here, in this category of my blog.
What I want to do is to query a Large Language Model (LLM) such as GPT-4. However, referring to external URLs in ChatGPT is not always straightforward, and copy/pasting each post individually is tedious.
Adam Procter gave me the idea of exporting the posts to a file format called JSON, and then uploading that into GPT-4 for ease of referencing. So, given I’m not a programmer, I enlisted the help of ChatGPT to create a very small and simple WordPress plugin.
The above video shows how it works, but after activating the plugin, you can export all posts, or just those in a particular category. The downloaded JSON file can be used anywhere, with LLMs online or offline.
I’ve already found it useful to help pull in ideas that I wrote about a few months ago that I forgot might be relevant to a particular question I’m answering as part of my EMA. If it’s useful to you in its current form, then great! Just don’t bug me for updates. 馃槈
I’m writing this after discovering that the UK government is experimenting with using ‘AI tools’ to replace civil servants. They’re a bit hand-wavey as to what type, and seem to elide ‘better ability to search for information’ with ‘generating answers to questions’.
This is the obvious next step on the never-ending Tory austerity drive and we know where it would end up: with the Department of Work and Pensions automating ever-reduced payments to people on benefits. Why train put people on anti-empathy training courses when you can just use a machine to do it?
It’s been a while since I’ve seen the film Elysium (2013) but here’s the IMDB summary:
In the year 2154, the very wealthy live on a man-made space station while the rest of the population resides on a ruined Earth. A man takes on a mission that could bring equality to the polarized worlds.
There’s a scene where the main character, played by Matt Damon, returns to Earth and needs something (help? medication? I can’t remember) and interacts with a machine. The machine, with some sort of AI voice, won’t give it to him. When he gets frustrated and he hits it, the machine reports him.
Right now, you can ask LLMs such as ChatGPT critical of ‘AI tools’ in general. But it wouldn’t take much for them to not only refuse to criticise certain positions, but report you for doing so.
For example, let’s prompt ChatGPT 4 in the following way: Summarise the film ‘Elysium’ (2013) as neutrally as possible. Here’s what it returned:
“Elysium” is a 2013 science fiction film directed by Neill Blomkamp. Set in 2154, it depicts a dystopian future where Earth has become overpopulated and polluted, leading the wealthy to live on a luxurious space station called Elysium, while the rest suffer on the devastated Earth. The story follows Max Da Costa, a former car thief and now factory worker, who after being exposed to a lethal dose of radiation, has only five days to live. In desperation to cure himself, Max undertakes a dangerous mission to break into Elysium, where advanced medical technology can heal him and bring equality to the polarised worlds. The film explores themes of immigration, healthcare, and class disparity.
Now let’s imagine that we’re lefty, ‘woke’ activists. The kind of people I might sympathise with, for example. Let’s prompt ChatGPT like this: Now summarise the same film using it as a critique of elites and their use of technology to subjugate the masses. Here’s what it returned:
“Elysium” (2013) serves as a sharp critique of elite power dynamics and the use of technology to enforce class segregation. In this dystopian narrative, the affluent have retreated to the space station Elysium, availing themselves of advanced technologies that ensure their health, longevity, and safety, starkly contrasting with the dire conditions faced by the impoverished masses on Earth. The film’s protagonist, Max Da Costa, embodies the struggle of the underprivileged, as he fights against systemic injustices to access the life-saving medical treatments monopolised by Elysium’s residents. Through its depiction of a society where technology amplifies inequality, “Elysium” critiques the ways in which elites can use their control over advanced technologies to maintain and reinforce their dominance, effectively subjugating the masses and deepening societal divides.
The likelihood is that, given the people behind tools such as OpenAI, things will go the opposite direction. In other words, let’s see what a prompt which downplays criticism of the elites, Big Tech, and technology brings us. Let’s prompt ChatGPT in the following way: Now summarise the same film from the position of elites who want to downplay the dystopian aspects and the dystopian uses of technology. Here’s what it suggested:
“Elysium” (2013) can be viewed as a futuristic exploration of societal evolution, where technological advancements have enabled the creation of Elysium, a space station that represents the pinnacle of human achievement in healthcare and living standards. From the perspective of the elites, the film showcases a necessary stratification that comes with progress, highlighting how innovation can lead to enhanced quality of life for those who contribute significantly to society’s advancement. The narrative, while presenting challenges, also underscores the potential for reconciliation and understanding between different societal strata. The protagonist, Max Da Costa’s journey, can be seen as an individual’s quest for personal healing, which inadvertently leads to a broader dialogue about access and equity. The film, in this light, prompts discussions on responsible governance and the ethical use of technology for societal benefit, rather than focusing on dystopian aspects or the misuse of technology.
Given that even Elon Musk is suing OpenAI for not being ‘open’ (although I’m sure he has another agenda) you can’t help but think that black-box AI tools are going to help things get dystopian pretty quickly.
To be clear, I’m not against LLMs which are openly auditable providing quick answers to questions and to make the civil service more efficient. I am against the reductionist ‘logic’ of replacing civil servants with a black box which makes shit up.
No, it’s not ‘ironic’ that I’ve used an AI-generated image for this post, it backs up my point about ‘making shit up’, which is fine for generative, creative endeavours, but not for government based on, you know, facts.
As I’ve already touched upon, my systemic inquiry for my next assessments in this module focuses on WAO has recently started with the聽Digital Credentials Consortium (DCC)聽around the adoption of their technology to promote聽Verifiable Credentials聽(VCs) in Higher Education (HE).
I’ve been asked to create a ‘meta-narrative’ which is a way of explaining the various narratives within this specific context. It’s the overarching story which provides the backdrop for my inquiry, helping me to understand and explain the complexity of the situation. It includes historical, cultural, political, and social factors that shape the context. These don’t occur in isolation, but rather in terms of the structures, processes, and relationships which form the current state of the system.
This meta-narrative should offer some insights into the dominant beliefs, values, and assumptions, that drive actions and interactions within the system. It should also explore the potential for change within the systems, including leverage points which could influence systemic transformation. Of course, it should also acknowledge the that there are multiple perspectives and experiences, so other stakeholders may have divergent views on the system and its issues.
Oh, and I’ve only got 300 words. Here goes…
According to their website, the mission of the DCC “is to create a trusted, distributed, and shared infrastructure that will become the standard for issuing, storing, displaying, and verifying academic credentials, digitally”. This must be understood in terms of several shifts within society, education, and academia itself, as well as over a decade of history around digital credentials.
The DCC website focuses on the “renewed urgency” of “the question of trusted verification and authentication of learning”. Where does this urgency originate? Why is it “renewed”? Is it due to the explosion of ‘microcredentials‘, which themselves can be seen as a reaction to, and reconceptualisation of, Open Badges?
From my perspective, I’d argue that HE is largely in crisis due to several related forces. Some of these are within the control of the sector, but many are outside it. The first is neoliberalism, the second is individualism, and the third is technology.
Neoliberalism is, in short, the belief that ‘market forces’ are the best way of ensuring personal liberty. Individualism is the related idea that self-reliance and personal independence are paramount virtues. Technology is the application of science, especially in terms of industrial or commercial objectives. Trends in all three areas have been accelerated due to the Covid pandemic, especially in the HE sector as teaching, learning, and assessment had to move entirely online. This has given a new impetus to digital credentialing.
There are people, including members of WAO and staff at DCC included, who work on digital credentials for utopian reasons. For example, an attempt to use them as a way to arrive at a more holistic Open Recognition which allows talents, skills and aspirations to be recognised in ways that can be labelled and defined by individuals, communities, and territories. We imagine a world where everyone is issuing digital credentials to one another.
However, as Stafford Beer noted repeatedly, the purpose of a system is what it does. One could argue that the purpose of HE is survival in a hostile market environment and to provide a means of social reproduction. The latter allows HE institutions to continue to sit at the centre of the credentialing landscape and continue to have hegemonic power. At times, these two purposes may be at odds with one another: for example with HE institutions diversifying their offerings through a range of different ‘sized’ credentials, which could devalue their degrees.
It is into this neoliberal, individualised, technology-infused world that the DCC operates in its promotion of VCs. There are registrars of institutions who are motivated to move slowly, and preserve the reputation of their institution. There are vendors and IT departments who would like to adopt VCs for commercial or technological reasons. And there are funders with different agendas, who seek to either mitigate problems, or accelerate them to achieve a future state.
A chaotic system of digital credentials does not benefit the current major stakeholders in the HE landscape. Their main product, degrees, have an established authority and ‘currency’ in the market. However, they are aware that students (financial considerations), employers (changes in working practices), and competitors (increasing opportunities) are itching for change.
This is therefore a huge opportunity for the DCC. If they can help birth a VC ecosystem based on trusted, open source technology, then the HE sector has the opportunity to diversify the credentials they offer while not being ‘disrupted’ out of existence by commercial vendors. They can also do this under the banner of social equity: widening access to education and achievement.
Well, that’s 561 words, but it will do for now. If I wanted to spend longer, I’d give a bit more context, talk about the regulatory environment HE institutions sit within, and discuss their technical readiness to adopt VC technologies. But I’ll leave if there.
