Is AI leading us to abundance or abundantly hard times?
Useful and better than human level AI for a large variety of tasks across different modalities is already on the mainstream curve. This won't be another 'AI is going to end the world' piece. The aim here will be to take a deep look at how automation leveraging AI is truly making life easier in some ways while equivalently exasperating social and economical pain in others. This phenomenon is effectively creating a 'pinch point' in modern society that is already impossible to ignore.
The advancement in the field of AI pushing towards developing true Artificial General Intelligence(AGI) following on to Artificial Super Intelligence(ASI) has been unrelenting. In fact the pace of advancement is more frightening than exciting at the moment with new papers being published daily pointing to improvements that trump or drastically enhance much of what was published and considered 'mainstream bleeding edge' just a few months before. There's so much more at stake here than another wave of job automation. Although, automation of large parts of an internet based economy is a great place to start.
Bad news first!
Five years ago would have people entertained by models identifying cats and dogs or capturing credit card or receipt details from a photo. Tools to enhance our every day life in small ways had arrived and were only getting better. Articles about jobs being lost to automation had folks thinking of how we had already seen this happen to the common factory worker through the introduction of robots. Although the message from industry experts on mainstream media was clear, progress is happening but your job is most likely safe if it isn't tediously repetitive.
Fast forward to right now in 2023 and we're looking at creative industries across multiple mediums in absolute shock. So much shock that the sense of disaster as a result of the disruption these latest generative AI models have caused has unified a people. Groups of people consisting of music and digital artists who are now seeking funding for lobbying campaigns to stop AI from taking their jobs. Programmers had a similar reaction when Microsoft trawled public code repositories on Github to train their 'Copilot' AI programmer models. Models are now able to create content and perform better at complex takes such as writing and comprehension better than the average human.
For the purpose of providing content as a way to pay the bills, digital artists and musicians, writers(fact and fiction) and literary editors(Grammerly for example) will be heavily automation based. In most cases removing the need for a human creator at all. If the last five years has completely disrupted the need for a majority of humans to enable content creation for a fraction of the price and time, what could the next half a decade look like? Film makers, education providers and even game creators are easily the next inline to have to face this tsunami of economical disruption.
This isn't just about the indie or gig economy based artist, programmer or digital musician having their services automated away, we're talking about the automation of entire value delivery pipelines. Think about generating film content, with actors that don't exist, music scores generated to evoke emotion at the right time, story writing and plot design, all handled by a set of AI's that are easily accessible. This could reduce a team of ten people in a particular department to one or zero in many cases. Being able to make a fictional feature film for the big screen at a fraction of the cost with a fraction of the people.
Google(and others) already on a clear path to AI video generation as this paper so expertly explained by Dr Károly Zsolnai-Fehér on his channel Two Minute Papers.
No job for the sake of making a living is safe and that's the point. Putting aside the lack of frameworks, both economical and societal, that would allow humanity to transition from a work based economy to that of an abundance based reality. Work for work's sake is a waste of human life, potential and time. Why create better than human level intelligence and not 'upgrade' a world economic system that needs intelligent work to be done as fast and efficiently as possible? While industry experts are pointing to 'manual labour' being swapped out for a more comfortable way of making a living, progress to replace human level intelligence sourced work continues with indifference. Meanwhile progress on managing the impact on humanity has had very little movement or investment. Regulation has always played a reactive role in protecting humanity from exceptionally disruptive technological progress.
The approach is usually response starting with a cliched 'never again'. In reality regulators are already restricted by the existence of bad or apposing geo-political actors. The reason for this already apparent to most, i.e the idea of the race to AGI and ASI already being the next generation arms race. Alongside the fact that new disruptive technologies are scaling at a global level faster than the previous generation of tech in the same areas. A great example more recently would be TikTok dominating the social media space and reaching a billion users in almost half the time it took Facebook, Whatsapp and Youtube to do so. ChatGPT reaching a million users in less than a single week! Restricting access to tech and platforms in one country doesn't necessarily mean others will follow as the economic potential is too great in a capitalist dominated world but more importantly the potential to progress humanity too!
So in the short term, it seems there is much pain to expect in the way of it becoming ever more difficult to exchange value for money in the current content creation and attention based economy that has economically dominated the online world for at least the last decade. Add to this physical work being automated away too and the picture starts to add that much more clarity as to where we're headed.
Yet there is hope, or is it 'Hopium'?
It would be understatement to say that we're potentially looking at a level of disruption we haven't experienced since the industrial revolution. It's almost impossible to understand what value humans would be able to provide aside from a human to human experience providing genuine connection. Think back to the 90's, if you're old enough, when the internet was on the cusp of exploding and changing our world forever.
No one would ever have thought you could build a long standing successful business posting personal product reviews or even envisaging the advent of the influencer industry or online gaming as a professional sport! Face-time doctors appointments and remote working, all a reality for millions of people in less than 2 decades. All this disruption with technology that only aspired to connect individuals regardless of geographical location alongside making information easily accessible. If the internet revolution felt like going from horse to car, the AI revolution will feel like going from horse to battery powered flying car.
Here is where the hope starts! The speed of progress reduces the time of struggle and suffering. Faster drug discovery, more efficient hospitals and medical facilities. Self diagnosis through the use of the internet has absolutely saved countless of lives and all it took was for the internet to allow practitioners to share online tools and questionnaires aimed for those searching about their aches and pains! Well a chat CGPT model could be enhanced to train on question, answer and diagnosis data allowing for a 'better than GP'.
This could revolutionise the issue of teacher to child ratio in classes as technically every child would have their own personal AI tutor...
What about education? Currently Ivy league level education and the highest quality educational tools are already completely free and easily accessible online. From open sourced recorded lectures to platforms like Khan Academy and Coursera offering content arguably better than most public schools and colleges. This has allowed millions to change career, improve their literacy, find new ways of expressing themselves or even pursue their true passion as a result of the reduced cost, convenience and high quality of information. A personal and private AI tutor could easily enhance the experience here and tailor learning to how the user responds to prompts and questions. This could revolutionise the issue of teacher to child ratio in classes as technically ever child would have their own personal AI tutor.
We find this one of the most exciting areas of how AI could impact society as we know it. Starting from the up and coming generations. There is a caveat here and that is censorship and limiting AI's according to what a company or government deems 'too' dangerous. We've seen this with Stable Diffusion and ChatGPT already in the way of both models being heavily restricted on responding to certain prompts including specific questions. Either way it can't be denied that this would absolute elevate the quality of education we currently having in most schools and colleges.
A not so balanced result
Ultimately the 'working for a living' lifestyle, we say, in the next 15-20 years from now will either be dead and a form of UBI introduced. Or along with this digital and economic revolution we just might land up with a human powered revolution and unrest that would make even the french blush. Right now as we're looking for where the frameworks that we have to help those who's lives have been completely disrupted through automation.
We can't all be programmers or technical prompters after a 3 month AI bootcamp! Not to mention that your new programmer job is also but the next GPT release away from being automated. Maybe it's GPT4, maybe GPT5, the point is that it's only a matter of time. It also stands to reason that better than human level intelligent AI for any industry means any services online or off provided by a human will sooner or later end up reducing that human supplied workforce by all or most of the current quota. The idea of this coming to fruition for the majority of value created by humans to be exchanged for money is becoming ever more a serious misjudgement of how fast AI models are progressing.
Without a form of UBI or alternative economical structures and schemes actually being implemented at a larger scale and made permanent after pilot successes - the near future is not looking to bright dear readers. With nothing but the 'Hopium' laced articles of more jam tomorrow until governments decide enough is enough on behalf of us all, or the people do spurring governments to act proactively. It seems we will have to wait for AI to really blow up the current economic model as this will be and is the easiest place for disruption to scale. We need more of these conversations to be had in the public if we're going to spur change and avoid more of the economical pain to come. This is our attempt, here's to a better future for all of humanity.