So much of our work (my work, anyway) can be described as “what we do in front of a computer”. What happens when we can talk with that computer and ask it to conduct those tasks rather than do them ourselves? What skills do we need to learn then?
This is the thrust of the Act-1 AI transformer. The example applications might seem rather modest against the more eye catching ChatGPT and Stable Diffusion examples. They have far-reaching and possibly more immediate implications, however. And they are only in the very early days of development. The business has a clear view of what AI can do for work:
“general intelligence is a system that can do anything a human can do in front of a computer”https://www.adept.ai/
Act-1 is something like a personal assistant, taking care of administrative tasks and also like a junior staff member, doing what we might call grunt work. No need to know how to use Excel to create and manipulate formulae, only the analysis we want to see. No need to search web pages to create that vendor shortlist, just state the query and wait for the results as the tool conducts the search and organises results for you. What will happen to knowledge work and knowledge workers as task automation creeps in to our daily efforts is going to be an interesting voyage of discovery.
Personally, having a robotic spreadsheet analyst to hand will be invaluable. Complex analysis is rare enough in my work that learning those deeper skills feels like a lot of investment for only occasional reward. That dilemma looks to have been resolved with this kind of tool. There will be many such skills that we will not need to develop any more. Understanding how to apply them and using good judgement in evaluating output will remain important. What workers need to be good at will change a great deal. It might also change quite quickly.
A significant shift in skills development priorities is a likely outcome of the adoption of these tools. Deciding on and defining clear instructions will be a more common need – clarity of objectives, understanding what information is available and articulating then refining needs will be foundational. Evaluating results and further defining needs will also be important. The skills to perform the tasks themselves will decline in value as automation becomes more prevalent.
The traditional route to job promotion may also narrow for new entrants. Those roles that worked as proving grounds for junior staff as they develop and demonstrate what they can do seem likely to decline. Or, the skills they need to show will change. What they need to learn will change.
L&D services will need to respond to the nature of skills development and also to the rapidly changing value of skills we are required to develop. What used to be significant will be less so, and that decline will be more rapid. Interesting times.
We will need to be alive to the potential automation of many tasks involved in content design and production. The work of ‘making learning’ (such a clumsy phrase) is starting to change profoundly. Workflow automation is a clear benefit, as is the creation of new content. Anyone can edit video right? Making stuff is getting easier to automate – and we do like to make stuff. Our most valauble skills will be elsewhere.