Digital First Mindset

There is a difference between having digital proficiency and having a digital mindset.  You can have digital proficiency and not have a digital mindset.  But it is really hard to have a digital mindset and not have digital proficiency.

A digital mindset is, per the article reference above, "a set of attitudes and behaviors that enable people and organizations to see how data, algorithms, and AI open up new possibilities and to chart a path for success in an increasingly technology-intensive world."

For starters, right now; most people are just hanging out and waiting for the future to arrive.  As is said, "The best way to predict the future is to create it."  I can tell you from first hand experience that being directly involved helps you understand what the future is probably going to be like.

Clearly something is up. I call what is happening "the great transmutation".  Others call it "the great upheaval".  Still others refer to what is going on as "the great progression".  My point here is that something is up and that something is big, big, big.  And while I am personally focused on accountancy, this change will impact pretty much everyone.

From my vantage point I see four fundamental mistakes that people are making that don't seem to have a digital mindset which could lead them to make very bad choices. In summary:

1. Digital is not software, it is a mindset.
2. Don't force computers to understand human oriented artifacts; create machine understandable artifacts and then also make those understandable to humans.
3. Algorithm intelligence is different than human intelligence.
4. The problem with artificial intelligence is not artificial intelligence; it is humans misapplying artificial intelligence.

First (#1), as Aaron Dignan points out; digital is not about software, digital is a mindset. The most dominant companies, no matter the industry, are digital-first. Think Netflix over Blockbuster or iTunes over Tower Records.  Aaron Dignan walks us through how we can have the right mindset to thrive in the future: We need a purpose, a process to support it, the right people, and (most importantly) these need to combine to make products that serve a community larger than any employee or organization. Dignan shows off plenty of examples and tells us what to adopt for our own work. “When we look at the companies that are really dominating, this is what they are doing.”  These companies recognize that "cyberspace" operates by different rules than "realspace". The "digital operating system" is the dominant paradigm now.  Interactions will be different. 

Second (#2), accountancy has operated under the same fundamental rules for about 7,000 years.  While the artifacts have changed over time the approach/process has not been able to change until now. Take one artifact as an example, a "report".  Reports have changed from human readable clay tablets, to human readable papyrus, to human readable paper, to human readable PDFs and electronic spreadsheets.  But now a transition has occurred and it is very possible for machines to actually understand reports to an extent.  There are two critically important dynamics that need to be understood here.  The first dynamic is the actual intelligence of these machines. Both over estimating and under estimating the capabilities of a machine will have consequences.  The second dynamic is that there are two approaches to machines understanding things can be employed:

1. Take the human readable artifacts (i.e. human first by design) and build software to try and make sense of that human readable information.
2. Reconfigure the human readable artifact to optimize it for effective machine understandability (i.e. machine first by design) and then also provide the functionality to convert the machine readable artifact into a human readable artifact.

People think that things like ChatGPT are short cuts.  They are not; they are dead ends for a lot of things.  They are excellent tools for other things.  To understand the differences properly, you need to correctly understand how deductive reasoning and inductive reasoning actually work. A craftsmen needs to understand the tools they are working with.  Personally, I am taking approach #2 and it is a lot of work.

Third (#3), artificial intelligence uses software algorithms to perform tasks that have been traditionally handled by humans with real intelligence.  Humans have their way of doing things; understanding the problems and performing those tasks.  Artificial intelligence, in some cases, can also understand problems, devise solutions, and perform tasks, projects, processes, and entire workflows. Computer algorithms, in some cases,  can deliver outcomes equivalent to or perhaps even better than human outcomes.  But an algorithm's intelligence does not go about devising solutions to problems the same way as humans intelligence would. This is not a good or bad thing; but it is something that is very important to understand. Watch that video above and contemplate how Amazon's warehouses now work.

Technology is very likely to alter many areas of accountancy and other types of professional services such as legal services and medical services.  Algorithms need data and information to work effectively.  That corpus of data must be structured into a format that is usable to the artificial intelligence.  That structure might be different based on the task you ask the algorithm to perform.  The proof will be in the pudding.  If something is not logical to a human; the probability that the information will be logical to an algorithm is remote.  To make algorithms good decision makers, those algorithms need good data sets.  Sure, technology breakthroughs can help make artificial intelligence useful.  But you cannot expect artificial intelligence to clean up a mess.  Humans need to clean up the mess.

Finally (#4), artificial intelligence can be problematic but not for the reasons people seem to think.  Sure, an AI software application can make egregious mistakes. But most mistakes will be made by systems designers that fail to construct systems appropriately.  Here is an example of what I mean.  Think of a self driving car.  What speed to you tell a self driving car to go?  We humans know what the speed limit sign says, but there are unwritten rules that the humans follow.  What speed will the designers of self driving cars pick?  Drive the speed limit?  Or, drive 10 mph over the speed limit, the unwritten rule we humans know which can get us where we want to go faster but still avoid a traffic ticket.  But if a self driving car gets a ticket; who is responsible for the bill?

Additional Information:

Third-Party Ethics in the Age of the Fourth Party

What is the Algorithm Economy?

All About Knowledge