The social contract and the inevitable transition from human-first to AI-first technology
I am covering two concepts in this newsletter - the social contract between corporations and autonomous technology, and the phases of change as we move from a human-first model through the current AI-assisted period to an AI-first world. I also cover how to plan for you own business and career.
The Social Contract
I attended Manchester University from 1990 - 1994, after two attempts at my A ‘levels, and read Accounting & Law, which was a 4-year joint honours degree. That was a good thing as all of my friends left after 3-years and I finally did some work. I left with a 2:1 and the R G Lawson Faculty of Law prize in Law & Economics and went on to become a qualified lawyer in the City of London.
My dissertation was on the Social Contract and how corporations were allowed to become legal persons that could contract, and even be convicted of crimes. This was due to the benefit they presented to society by bringing together the units of production to produce goods and services at scale. I won’t go into the writings of great philosophers such as Rousseau, Locke and Hobbes, about the state of nature and individuals surrendering their individual rights for the good of the many.
The point of my dissertation was that the recognised stakeholders of corporations widened over time from the mill owner to shareholders, to employees, and eventually to the global environment. How could all of the public costs (e.g. pollution) of doing business be captured in the private costs of corporations and priced into products and services? These phases or recognition are still work in progress since 1600 when the East India Company was granted its Royal Charter and the arrival of limited liability companies in the mid-1850s.
Our ‘Ted’ moment
f you have seen the film ‘Ted’, you will be aware of the legal complexities of a sentient stoner teddy bear being recognised as a being with the same rights as humans. Our laws and regulations are usually 5 – 50 years behind commerce and societal mores. In the case of exponential change, they have little chance of keeping up. Companies will soon be employing artifically intelligent beings as knowledge workers, followed by robots designed by AI to take on physical labour. Within the next 50 years we can expect not just artificial general intelligence (AGI), but also artificial super intelligence (ASI), which will essentially be a new species that is orders of magnitude more intelligent than us.
Preparing for change
Has society given the go-ahead as it did with corporations? Are we ready for a ‘Ted’ moment? What is clear is that we are accelerating through phases of evolution into an age that will impact us all. Yes, we don’t know when this point of no return/event horizon/singularity will be, but it's coming. Whether you run a business or have a career that will be, or already has been, impacted, you need a plan.
At the macro-level society needs a plan as AI will be far more disruptive than limited liability companies. Autonomous companies or Decentralised Autonomous Organisations (DAOs) running on block chain technology may not require a board of directors to think and make decisions.
Sam Altman, the CEO or OpenAI, which launched ChatGPT wrote, not unsurprisingly with the benefit of hindsight, a prescient blog in 2021 about the future of our economies, work, society and fiscal policy. He proposes that Moore's Law will apply to everything - costs will fall as output increases in an ever-accelerating process leading to exponential change. His proposals would require a new social contract and even new nation state constitutions.
From human-first to AI-assisted to AI-first
The inevitable transition
Sam Altman’s blog sets out an inevitable technological future, although I would question his thoughts on how to generate Universal Basic Income (UBI) for the millions (or billions) or people that are no longer needed as units of production.
What are the phases of change and what can you do to stay ahead?
The chart above is a take on several others, but I don’t believe that humans, who are used to slow, linear change and knowledge sharing through slow academic processes, will ever be able to estimate or even understand the pace of change that we will experience.
There are three phases, although robotics will take a little longer - or at least until AGI or ASI starts to design and produce them in ways that we can’t begin to understand. Unlike humans, as soon as one networked AI solves or discovers something, every other AI will know immediately and build upon that new knowledge.
The cost of labour, knowledge generation, production and content are going to zero. One of the first enormous technical advancements will hopefully be nuclear fusion which means the cost of energy will follow.
Stage 1: History was human-first:
From the dawn of time, humans have utilised beasts of burden, tools and machines to enhance our limited physical ability, to learn and to speed up production. It wasn’t until the mid-1700s that the agrarian revolution increased the efficiency of farming to release labour from the fields. They found work in the mills and cities of the industrial revolution. Whether consumer electronics, personal computers, smart phones, or the emergence of the internet, technology has made humans more productive and more efficient. Economies have grown and millions of new jobs have been created.
AI is very different as, although it will generate exponential growth and abundance, it will eventually displace almost all human involvement in the production of goods and services.
Stage 2: Today is AI-Assisted
Disparate research into different types of AI – from computer vision to text-to-speech – has now coalesced and resources are now focused on Language Models (LLMs) and Generative Pre-trained Transformers (GPTs). There is other deep tech research being conducted by private companies and governments. The realisation of the enormous value of, and interest in, the recently launched platforms has kickstarted a technological race between companies such as OpenAI and Microsoft, Google, Amazon and many others. It is also a race between nations to create a powerful economic weapon, as well as destructive weapons. including computer viruses that are not just designed by AI - they are AI.
We are already AI-assisted and always will be from now on. Our children will may just remember the world pre-AI, just as I can just about remember a pre-internet world. During this period many knowledge and creative workers will become either far more productive, or unnecessary. In 1972 the simple ‘Pong’ tennis game was launched.
I was born in 1971 and it has taken my lifetime (50 years) to get to Forza Horizon. We are now being assisted by ‘Pong AI’ and it will only take a couple of years to reach to the equivalent of 'Forza AI', and the pace of development will only accelerate.
50 years from Pong to Forza Horizon
In the next 2-5 years, humans will be increasingly assisted by AI and the technology will impact a diverse range of industries and roles although creative and higher-paid knowledge worker roles are at just as much risk as administrative ones. A company can expect a 30%-50% productivity improvement in most knowledge worker roles.
We can also expect the deployment of more 'simple' robots, such as the thousands working 24x7 in Amazon and Ocado warehouses. Robots cutting your hair, caring for children and building walls will take longer due to the costs, investment payback period and rate of evolution. AGI or ASI would reduce the time to robots that can do anything a human can do.
Stage 3: AI-First
This is the future described by Sam Altman in his blog. There will be no job or activity to produce a good or service that will not be done by AI, or a walking and talking AI. Humans will need to find something else to do as the concept of a job will no longer be relevant. The emergence of even more powerful AGI, or even ASI, will result in a profound acceleration in the capabilities of robots.
In this world we will need to be aligned with AI and eventually a species of ASI. We will need revised notion of the Social Contract and new laws before we hand over the reins, but it will no doubt result in the usual too little, too late.
When will you or your business be impacted?
During Stage 2 (AI-assisted) the main risk is your competition making better-use of the AI available today and having a flexible approach to deploying improved AI versions over time. The same goes for an individual who can now be far more productive.
Use his high-level list to guide you:
1. Digital vs. Physical:
Analyse the extent to which your offerings embrace digital solutions as opposed to physical products or services, although many software-as-a-service (SaaS) business models will be disrupted by natural language LLMs capable of understanding and completing complex tasks.
Businesses creating digital products or services will be impacted faster than those creating physical product.
2. Knowledge Work vs. Manual Work:
Assess the reliance of your business on knowledge workers versus manual labour, as the former will be disrupted faster.
Recognise the evolving role of AI in augmenting knowledge work to increase productivity and the impact of future hiring (in May 2023, IBM announced that it has halted recruitment for 7,800 jobs that are likely to be automated).
3. Repetitive Work vs. Custom Work:
Evaluate the balance between repetitive tasks, especially at a computer keyboard, and higher value, or custom work.
Anticipate how automation and AI can streamline repetitive processes, allowing human resources to focus on more intricate, creative or personal tasks that add value.
4. Regulated vs. Unregulated Industries:
Consider the degree of regulation in your industry and the potential implications for AI adoption and the speed of acceptance.
Stay informed about evolving regulatory frameworks to navigate compliance requirements effectively.
5. Commodity vs. High-Value (Luxury):
Reflect on whether your product or service falls into the commodity category or offers unique value and luxury.
Identify opportunities to leverage AI to differentiate your offerings and provide personalised experiences to customers.
Embrace AI and optimise your business now!
The rate of change is difficult to estimate and even hard to imagine, but change is coming and catching up to agile competitors, that are already leveraging the power of AI, will be difficult.
Some argue that LLMs are already more efficient than the human brain as storing, recalling and assimilating information with a similar number of connections (neuron equivalent) in ChatGPT-4. At some point in the not-too-distant future, AI intelligence will surpass human intelligence, and I am not even talking about AGI or even ASI. Teams of connected specialist LLMs and GPT agents will collaborate to complete tasks.
Two weeks ago, I installed AutoGPT on a computer to test this model and it is an interesting albeit over-hyped demonstration of what could be possible. This week OpenAI launched its ‘plugins’ to a range of external services (e.g. flight booking), which is a productised version of AutoGPT. This is how fast the market is moving.
Do you want to learn more? 'AI for Business - Boardroom Update'
I am taking part in an event on 25th May in London with MIA Cloud Academy:
'AI for Business - Boardroom Update'. Get the AI insights and advantages for business you need to know about. If you would like to join us, DM me for a discount code. Learn more here.
I have also cofounded a new business to support small and medium-sized business to understand and implement AI.
In fact we decided to call it Implement AI. Sign up for updates here.