AI will come for knowledge work, 3 ways we might respond.

AI will eat the world

Given enough time and compute resources, AI will replace 90% of the knowledge work we do today. At the current rate of automation that time may be shorter than we think. The 90% that is replaced by AI will be the simpler things, repetitive well defined tasks. This replacement is inevitable, all we can do is decide how to respond. We can choose to embrace the change and let that 90% open up a world of possibility or we can fight it until it wins and be left in the dark. At a political level we need to prepare for this change and ensure we take the steps that leave us all better off rather than leaving significant portions of the workforce without employment. When this replacement happens, its not the end of the story. This trend will continue so long as replacing knowledge work with AI provides an efficiency gain. A cycle that will likely get faster as we progress.

Why it's inevitable that AI will take over most knowledge work

As individuals we crave novelty, as corporation we crave efficiency. In both cases the work we do continues to become more refined and repeatable. In doing so, we can work faster and stretch ourselves to work on more complex tasks. The work at the bottom of the pile quickly becomes mundane, we look for ways to do that quickly and hopefully spend time on the 10% at the top. Economically the knowledge work in the 90% is ripe for automation. When tasks become well defined they are strong candidates for replacement by AI.

The final 10% remains mainly because it is new. It is knowledge work primarily focused around discovery and invention. Often hard to describe this work sits at the limits of our understanding of whatever task we're working on. That however, is not a permanent thing. As we progress and repeat work at this level it quickly becomes systematized. The top 10% shifts a little higher and the cycle continues. This shift is completely disjoint from AI. Today, when this cycle happens we create new roles and move work to new roles. When AI (or any complex automation) enters this loop, the work can be offloaded at a significantly reduced cost. This will either create an opportunity or a threat, depending on who is in charge.

How governments could respond

Generally speaking, governments can take three distinct paths in reacting to this change. First, they can bury their heads in the sand. This is usually in an effort to make it the problem of a future administration. Secondly, they can actively work to prevent it. This can be a hostile approach, outlawing any AI or a combination or regulatory capture and policies that make the use of AI a great burden. Finally, and perhaps ideally, governments can embrace the coming change. This is the most complex path and must include designing an economy around this change that does not lead to mass joblessness. The 90% will be taken one way or another. The path chosen primarily determines what happens to the remaining 10% and whether it grows to bring new value or is captured by a minority who hold power.

Option 1 - Ignore the signs and kick the can down the road

Denial is a key stage in a lot of human changes. Denial in this setting often comes across as a dismissal of the potential capabilities of AI. I see this often in my work in health AI. Early discussions amongst clinicians started with a denial that AI could possibly perform clinical tasks. Luckily for the clinical world, this stage passed and a human-AI future is being designed and accepted. At the larger scale ignorance of AI and its potential to replace human knowledge work is a threat to current and future citizens.

No economy today is designed to thrive with high levels of unemployment. An ignorant government will attempt to draw focus away from the threat to more pressing issues. The challenge for government is the balance between immediate and long term threats. Whether it is climate change or AI driven unemployment, inaction today will likely amplify the problem in the future. While government should not need to create these future jobs, an active dismissal will leave the private sector unprepared. While innovative enterprise will adapt, this policy hurts those most vulnerable most of all.

Option 2 - Active blocking or rejection

Governments can, both aggressively and through poor policy, work to block major innovation driven change. An example of this from recent history is the tragic end to the Australian automotive industry. Government policy was set to focus on maintaining the status quo and keeping an industry alive long after it was viable. Poor policy can also appear as overzealous regulation. Similar to ignorance this tends to affect smaller players more heavily. Large firms have the resources to gain regulatory capture and block competition. Without accessible access to AI, especially in knowledge work driven industries, we'll see power condense in a small handful of players and lose the benefits of an open competitive landscape.

On an international scale this also opens a nation up to losing its competitive edge or becoming dangerously reliant on the import of knowledge work. The pandemic and current geopolitical events have opened the eyes of many to the need for energy security, food security and a local medical manufacturing base. In a similar vein if a country relies on the import of knowledge work around engineering, medicine or other services it leave itself vulnerable in the event of a crisis. Given the ever decreasing capital requirements for AI this further emphasizes why rejection is likely the worst course of action.

Option 2 - Embrace, prepare and thrive

This is the hardest path. At least in so far as it is the hardest path that leads to victory. Preparing for a future where 90% of knowledge work jobs are gone is complex. Governments role in this is to provide safety to those it may displace and encourage reskilling though policy and taxation. That does not mean direct intervention or job creation to replace that 90%. The best path a forward thinking government can take is to create an environment in which the remaining 10% can expand into new industry and provide both jobs and an international competitive advantage.

Option 2a is one to be wary of. I've have often seen efforts to support changes like this paid only lip service. Special interest groups are made and spend years creating reports and strategic plans that give the public some peace of mind. Many of these efforts go on to waste large amounts of time and money in the process. To truly embrace the changes, governments will need to make concrete policy changes. These changes won't be created all at once and will require adaptation to adapt as the workforce changes. This need for ongoing policy making for a new and complex technology is no easy task. As citizens we need to make sure those in charge have the means and the desire to see these changes through.

What we do with the 10% that remains

Assuming we choose to embrace the coming change we need to work to decide what happens to the 10% that remains. The darker path is one where the 10% is controlled by a small group of corporate entities who hold some capital advantage over the masses. This outcome holds a strong chance of rebellion and unrest or worse, needless bureaucracy. The dystopia where our days are filled with needless busy work to keep us paid and buying more for no other goal than to repeat the cycle. A future that feels a little closer every day.

The optimist in me sees a world where the free time pushes us to new heights. Where experts like clinicians are no longer scarce which allows top quality healthcare to reach millions more people. There's also a path which means we all do a little less. Five day weeks become four day weeks and we spend more time on a 10% that is more bespoke. Pursuing art and hobbies that serve only us and those we love.

Subscribe to Elliot C Smith

Don’t miss out on the latest issues. Sign up now to get access to the library of members-only issues.