Artificial Intelligence & Human Endeavour III

Part III of III: AI, Useful Tool or Ascendant Overlord?

We left you at the end of part II with a note about the intersection of AI (and economics) with moral, ethical and social considerations. The central concern a large number of people have with AI, is whether it will inevitably supplant humans in many aspects of work or everyday life in much the same way that machines captured manufacturing industries. We don’t think so, at least not entirely.

In attempting to address at least some of these issues we’d like to start at perhaps the most drastic and frightening or hopeful, depending on your perspective, potential AI outcome: replacement of human interactions, guidance and companionship. In the last few months, the media has been flooding readers with articles on people who have variously turned to AI as counsellors, friends or lovers, with varying outcomes. For some, AI has alleviated feelings of isolation, provided a sense of belonging, given them purpose, or helped them to overcome setbacks. For others AI has precipitated financial losses, professional implosions, physical and social isolation, sextortion and other forms of abuse, self-harm, and suicide.

The desire for human connection and communication is something we have mentioned to repeatedly in this series and if that comes in the form of AI, it is just as welcome and helpful as a corporeal human being for many people. This is not so different to domesticated animals that have provided companionship and comfort to humans in ways that are acceptable or even encouraged in most cultures. Many see AI as a technological extension of human-animal type supportive interactions. We don’t have a value judgement either way, except that we believe everyone has a right to dignity and happiness in life, as long as it doesn’t come at the expense of others, and if AI provides that, it can be force for good.

Despite this, concerns about the development and usage of AI are apparent, and justified. While animals may provide some level of interaction, they cannot directly communicate in human language to prompt, suggest or guide an individual, but AI can. This prompting or guidance by AI is highly variable and subject to fairly extreme changes when the program’s code is updated, including updates that occur for reasons that are not transparent. Essentially AI is a thing subject to sudden amendments and changes at the whims of it owners. There’s a very similar situation in the pharmaceutical industry where many drugs and therapies have indisputable, even life-saving benefits. Those benefits are a side-effect, or at the very least a co-effect, of the pharmaceutical industry’s drive to make profit with their business models predicated on greater work through perceived benefit, the benefit being human health and wellbeing. Again, at best this can be a symbiosis, at worst the parasite will sicken or kill the host. The same principles and caveats apply to AI companies, including OpenAI that has recently moved away from not-for-profit and public good directives, and into the for-profit space. Therefore, it behooves regulators and legislators to pay careful attention to this new industry, and create the guardrails the public needs for the greater good. These regulators and legislators could do far worse than starting with Asimov’s Laws of Robotics.

We believe that a for-profit motive can exist alongside public good works, but these must be guarded by sufficiently broad, and in some cases punitive, laws that codify ethical standards. Which brings us to a less existential, but still critical consideration of human endeavour and the human condition, namely: AI and the supplanting of human employment and endeavour.

Some of the very first points made in both part I and part II of this series revolve around feelings of a loss of primacy, and part of that is the loss, reduction in or redefinition of employment. It is probably inevitable that some redefinition of roles and tasks will occur, but we don’t think that AI will entirely replace humans in most cases. Perhaps the best chance we have to integrate AI is right now, while it is incapable of creating original thought, ideas or material. Although it can amalgamate and summarize things these basic tasks could be deployed alongside advanced efforts. Probably the strongest reference point here are Artisans. This term is usually associated with luxury goods and services from so called “one percent industries” that include shipbuilding, fine arts, apparel, jewelry and accessories. These are examples of artisanal work or sectors that revolve around scarcity, high-level craft and elevated, at times astronomical, prices. But artisans are not confined to the luxury space and their contributions and influence, is as present and relevant at your local grocery store, where you have the choice of multiple brands and types of products. Many of these are produced to address a specific need, want, or preference, and come in a range of prices reflecting quality, preference, marketing, and many other intangible considerations. We contend that: 1. those purveyors have identified their niche market and set out specifically to communicate with, and serve it and 2. They provide excellent products and service in a crowded marketplace, meaning they are able to be recognized and paid for their efforts whilst remaining competitive. Some of these artisans have swapped [high] price-points for volume, although the price to sales balance (and therefore effort and income) commonly exists on a spectrum, rather than at a threshold. These products and brands (and therefore, their artisannal producers) are flourishing in towns, cities and regions all across the globe, alongside mass-market basics from industrial manufacturers and suppliers. When the pace of manufacturing sped up during and since the industrial revolution, these skilled individuals found a way to coexist, and thrive.

The pace of technology is still accelerating. ChatGPT was only widely released in 2022, and fundamentally changed many things as we know them, to varying levels of delight and horror. Within weeks there was a term coined for unoriginal and low-quality ‘writing’, ‘art’ and ‘commentary’ produced by AI:

slop

Currently enjoying a resurgence, slop is a fairly old word in the English language, likely mid-16^th century in origin from Middle English and German, with various iterations and meanings as an improper noun, verb and adjective. None of slop’s historical or current meanings have been, or are, especially complimentary. There is recognition that those using AI exclusively produce slop, and those shunning AI still have to deal with it. Similarly there is recognition that products and services produced by corporeal human beings can be creative and original, while slop produced by AI, no matter how compelling, cannot.

Individuals that participate in creative arts often experience intense competition but no more so than, for example, research science or medicine. Within this paradigm, AI may force individuals to be more responsive to the world around them rather than relying solely on their internal compasses or motivations in much the same way computers aid, and force, doctors to prescribe more accurately and appropriately. But even allowing for this competitive and external focus, the solution is not as simple as saying “raise your game” or “do better” in isolation, because there will always be individuals for whom “good enough” is, well … good enough. Probably including many that have adopted AI wholesale. There’s little argument that when AI is wildly inaccurate, such as when it makes up citations for legal briefs (also called “hallucinations”), it is substandard. But where do substandard and good enough collide? This is as hard to answer as where the boundaries between good, great and exceptional lie. To take a particular stance, we believe that AI has a place as an aid to human endeavour, not a substitute for it. Therefore, injudicious use and lack of oversight by real people are as much to blame for substandard AI output as the computer algorithms themselves.

Since we can’t put the genie back in the bottle, perhaps the only way to ensure AI is an adjunct not an overlord is to demonstrate the superior benefits and value of human contributions. Casting aspersions on AI’s users and purveyors will simply sow anger and resentment in a large part of the population. In contrast, demonstrating human excellence may produce increased appreciation, harmony and opportunity, which is surely the better option in a time of increasingly fractured relations.

To co-exist with AI, harness its power and create opportunities rather than becoming overwhelmed, resentful, or overtaken by it, maybe we all just need to keep creating. And shine.