Playfully Uncovering AI's Serious Novelty
Let's treat AI as AI to discover new opportunities found in its unique characteristics.
Here is the updated version published in Tech Policy Press on the 11th of June.
Tl;dr - AI can and will be applied as a direct replacement for many existing processes in the name of efficiency. However, we already see behaviour from AI which exceeds understanding and prediction. Appreciating and focussing on this unknown, non-human behaviour will provide opportunities for novel experiences, more applicable policy, and AI applications currently outside our imagination. This post suggests three ways to elevate the potential for you to notice these opportunities.
AI is receiving a lot of hype right now across many non-tech networks (including my octogenarian grandparents) with a variety of opinions about potential applications, risks and viability.
There is often a ‘bubble’ like quality to these opinions with some being repetitive, inflammatory or (perhaps purposefully) misinterpreting features of AI's development - a selection of which manage to do all three... This occurs on both sides of the debate and seems to draw discourse towards a cocktail of the loudest pundits’ most shocking dystopic or utopic sci-fi analogies and away from what I believe is a more stable, considered and inclusive development of AI .
This is not intended as an AI apology or blind optimism (but I truly welcome your comments). Rather, this is an attempt to refine, combine and communicate the exposure I have been fortunate to have with many of the leading voices in AI's design, application and governance over the past months of my fellowship at ToftH.
I hope this post will complement your existing AI knowledge (of any level) with three suggestions and a few examples for how to productively engage with this often amorphous and complicated topic.
To specify what I mean by AI in this post. I use examples of generative AI, large language models (LLMs) in particular, since this is behind the recent, popular and accessible AI applications, like ChatGPT, Dall-e, Midjourney and Stable Diffusion (plus a few more I will mention in the final section). However, the suggestions apply to deep learning approaches more broadly.
Who is this for?
This is for anyone who wants a rough framework with which to assess the developments of AI and know when they’re getting dragged into other (likely very interesting) tangential discussions. More specifically, this is for (nearly) anyone who is:
Wanting to have novel experiences using AI.
Building products by applying existing AI models - to highlight how to escape obvious approaches.
Investing in AI - to help identify what currently unnoticed futures will be afforded.
Making policy - to get a clearer view of AI and its novel implications rather than seeing it as a continuation of existing technologies.
Why is talking and thinking about AI so difficult?
Simply put, we are not treating AI as AI. Rather, many discussions seem to force it into existing historical categories of very hard social issues that have been raised and repeated for hundreds of years and come with their own specific baggage. Most of these issues are very interesting with very inspirational people working on them. However, I believe these issues cannot be solved by reducing them to the application (or removal) of AI.
Three ways to notice novelty, develop a more productive approach and avoid common conversational treadmills.
I have attempted to find the common ground between the opposing sides of the AI debate. Using these I have developed three suggestions to provide a means for you to identify common tangents, avoid getting dragged into them, and notice the unexpected novel elements of AI.
1. Recognising the importance of but not reducing AI to power dynamics.
Perhaps the most popular framing of AI is just another means to consolidate or distribute power. Whether that be the embedding of social norms (and crucially discrimination), enormous compute requirements, access to training data, geopolitics, or potential effects on labour. These are worthy conversations to understand and work on social dynamics. However, the issues discussed can be found in writing over the past ~200 years and are almost never novel to AI or provide specific insights into it.
One of the most prevalent discussions is the social effects of automation, which to me, sounds like mutated forms of critiques found in places like Benjamin's ‘The Work of Art in the Age of Mechanical Reproduction’, Marx's ‘Fragment on Machines’, Engel’s ‘On Authority’ (which takes the exact opposite view to Marx) and Capek's ‘R.U.R’. to name just a few. A VERY reductive thematic summary across all of these is “there are tasks which humans like/should be doing and automation will make this more/less accessible”. As a demonstration, how similar does this quote from Capek’s R.U.R., released over 100 years ago, sound to today’s automation optimists…?
“…within the next ten years [Robots] will produce so much wheat, so much cloth, so much everything that things will no longer have any value. Everyone will be able to take as much as he needs. There’ll be no more poverty. Yes, people will be out of work, but by then there’ll be no work left to be done. Everything will be done by living machines. People will do only what they enjoy. They will live only to perfect themselves"
These authors are all (mostly) interesting, providing a backdrop and conceptual history for what we are experiencing today with AI. To focus on them however, ultimately grounds the conversations in the past and expands it so much we miss the specificity that is crucial to understanding what anyone can do about the challenges faced. These are part of a broader conversation about who/what we trust to take on activities that have historically been considered exclusively for humans and what the social implications may be.
To contradict myself a little, here is a small tangent if you do want to discuss AI as relating to power…
Automation as its own topic does matter so I would recommend two books from the past few years. Daniel Susskind's ‘A World Without Work’ provides a modern take on this issue including the potential and risks of Universal Basic Income. Azeem Azhar’s ‘Exponential’ covers a broader range of topics but has a good point on successful organisations growing their headcount with automation but ‘bad’ ones just following trends without finding new roles for staff.
Also, given AI does have implications for power dynamic and the political, how can we know when it is suitable to view it in these terms? Langdon Winner’s article ‘Do Artifacts Have Politics?’ is a useful guide when assessing any technology in this way. To quickly touch on this (but I would recommend the whole article), there are four means (as refined by James Plunkett) to review any technology’s political implications: its use, its default beneficiaries, its compatibility with political systems, the required socioeconomic conditions for it to thrive.
Ok, let’s get back to the point: AI can be used for the gain and consolidation of power and many people (including me) want to and should work on this. However, we must acknowledge this and then look beyond to find what novel applications and experiences are possible and what sort of new institutions we need to take advantage of and provide protection from them. You can find examples of what this novelty looks like in the final section.
2. Questioning assumptions about 'human characteristics' and decoupling them from AI’s assessment.
So, in an attempt to cheat/narrowly avoid trying to answer the question "what does it mean to be human?" I will assume we can agree that there is no universal answer. If this is the case you'll likely see that the effort to compare AI to humans as a collective quickly seems to lose its meaning as there is no meaningful centre or ‘average human’ with which to make a comparison.
Do all humans think in the same way? Can all humans see? Can all humans walk? The list goes on and to me (and many others), that is a great thing - making space and exceptions for differing modes of interacting with the world not only increases inclusivity but also provides opportunities for novel and potentially beneficial perspectives to flourish. Just two examples of contemporary authors writing about this broader topic are Caroline Criado Perez and Matthew Syed .
Implicitly however, we still seem to develop towards our own type of intelligence. Many tests (including the Turing, Coffee, Robot College Student, Employment and Flatpack Furniture tests) and lists of assessment criteria for artificial general intelligence (AGI) are being created by and based on universal ideas of… human intelligence. Differences in intelligence is not a novel point to raise and in the past, I have seen it applied to the differing learning capabilities of children - famously depicted in a metaphorical cartoon assessing animals' ability by asking them to climb a tree.
Phew, was that another tangent? Let's bring this back from broad social issues to thinking specifically about AI. We are seeing a disjunction of things we previously thought had to be grouped together for intelligence and so must avoid reliance on outdated human-centric measures (e.g. Is it conscious? Can it pass the Turing test? Does it understand the meaning of its own output?).
Chat GPT, Google's Bard or Bing AI demonstrate extraordinary speed and abilities to replicate writing styles but fail miserably at showing a conventional understanding of the topics or how they obtained the ability to respond (for more on this, check out Rohit Krishnan’s recent piece). This may not be a lasting quality of these large language models (LLMs) with products like Perplexity seemingly able to provide the verifiable context we desire. However, by fitting LLM chatbots into our current (perhaps questionable) expectation that when we ask a computer a question it must give us ‘truthful’ reliable information we miss the playful opportunity to find new ways to interact with this novel non-human technology.
For some, this non-humanness is seen as something to be feared with a popular (in specific Twitter circles at least...) cartoon showing GPT-3 as a "weird alien" with a happy mask as a metaphor for the human-friendly interface included in ChatGPT developed through Reinforcement Learning from Human Feedback (RLHF).
Whilst I like the playfulness of this depiction, I think it is misleading about our view of and relationship with AI. I believe the fear and perceived danger of AI is not in its form or current function but in the expectations we have of it and consequently where we are willing to apply it. We fear reckless implementation in the name of the (human) implementors' desires.
This sounds a little bit like reducing AI to power, doesn't it? I'll let you explore that topic with the links already provided and focus more on avoiding using humans as a benchmark for AI. What the alien cartoon does show us is a novel type of intelligence/being/thing which we can investigate and over time perhaps understand novel applications to existing problems, behaviours, and experiences we did not anticipate.
3. Appreciating how different the development of AI is compared to other technologies.
The final approach I'll mention is perhaps the simplest and one I have heard several times from those developing AI. Deep learning in particular is visibly different from all previous technologies. I don't mean that in a clickbait-headline type way but as a means to distinguish it from the history of engineering as a skill to build machines which do not deviate from their intended design. The ‘mistakes’, ‘hallucinations’, and current impossibility to fully understand why a system came to a particular conclusion highlights this novelty.
In a recent conversation, a leading figure in the development of foundation models (the basis for much of the current AI development since 2018) suggested we ought to think about AI more like a biological system (but without extending this analogy) where we discover features rather than in an engineering design, such as a bridge, where one must know the outcome and function before building it.
The point of this biological framing is to move away from a desire to optimise for output and accept what may seem like mistakes as essential (but not always relevant) ways to understand and use these models in novel ways. For example, two models may produce the same output but the way they get there can be completely different and provide no means of knowing which will perform better at seemingly related tasks.
Appreciating the difference between AI and existing technologies seems particularly important when discussing AI policy (for both its stimulation and mitigation). Insights from existing technology will miss novel risks and opportunities by focussing on those known to us. Another (lightly edited) cartoon, this time of the streetlight effect demonstrates this type of issue well.
We are faced with big questions such as what data can LLMs be trained on? What are the copyright laws for images or text outputs? Can we trust 'facts' stated by LLMs? None of these I believe can be answered effectively by treating AI purely as a product of traditional engineering design. Searching for existing regulations beyond those for technology seems like a good place to start.
Practical ways to experience this novelty
I hope and expect that using the three suggested approaches will open up some new ways to understand and use AI. To me, it has become clear that generative AI applications already hold the power to enable some experiences beyond existing technology. Removing the expectation that we will be given perfect finished products and appreciating instead the novel output beyond our initial desires, in my perspective, perhaps the most liberating and exciting shift.
To give an idea of what I mean, here are some examples of popular applications I've been using (beyond ChatGPT and DALL-E) that take 2-5 minutes:
Generating dreamlike videos based on text prompts with Deforum.
Turning your sketches into more detailed and stylised images using text prompts with Scribble Diffusion
Creating reactive visuals for music of your choice based on novel-generated images within a particular theme with WZRD.
Mutating or combining existing photos to create new images with Midjourney or videos with Genmo.
Asking questions of Perplexity to get a summary answer with several links for further reading.
Over to you
I wish I could point at exactly the type of novelty I hint at throughout this piece and I am currently trying to build products towards this goal. I would love to continue this conversation and get your perspectives so do comment or connect. But for now, please use this post as a rallying call to take a slightly more curious and playful approach in developing what may be one of the most serious and impactful technological developments of human history.