In modern business processes, the adoption of generative AI has moved beyond mere tool utilization and is now entering the stage of systematic integration.
Beyond this lies a new era of intelligence: "Symphonic Intelligence."
This article will explore the current state and future prospects of generative AI utilization from two perspectives: iterative work and flow work.
Iterative Work
In a previous article, we analyzed the perspectives of "iterative work and tools" versus "flow work and systems" as viewpoints for enabling generative AI to perform business tasks.
Iterative work refers to tasks where humans semi-unconsciously combine multiple distinct concrete tasks and proceed through trial and error.
And for this iterative work, tools are optimal. By selecting tools that fit various tasks, work can be progressed efficiently. Therefore, it is necessary to assemble the required toolkit and become proficient in its use.
Currently, when generative AI is utilized in business, the vast majority of cases involve generative AI as a tool.
Much of the discussion about improving business efficiency with generative AI almost always refers to adding this new and powerful tool to the existing toolkit humans use for iterative work.
The Problem with Iterative Work
On the other hand, as pointed out in the previous article, the efficiency gains from tools in iterative work are relatively limited.
As tools become more efficient, humans eventually become the bottleneck. Ultimately, the barrier of human working hours cannot be overcome.
Furthermore, there's a significant gap in the efficiency and accuracy of iterative work between veteran employees and new recruits, and it's difficult to bridge this gap. Therefore, even if you want to double the workload next month, you can't handle it without people with veteran skills.
To resolve the problem of humans being the bottleneck, it ultimately comes down to replacing everything with artificial intelligence.
However, current generative AI does not yet possess that level of performance.
Moreover, even seemingly simple iterative tasks, when examined closely, consist of a large number of unconscious tasks.
For this reason, they couldn't be reduced to conventional IT systems or manuals that anyone could follow, and thus relied on human proficiency.
Unless these numerous unconscious, proficiency-requiring tasks are organized and the necessary know-how for each is codified into knowledge, generative AI, no matter how much its performance improves, cannot replace human work.
Transforming into Flow Work and Systemization
To address the goal of distributing tasks within the current performance limits of generative AI, and the goal of organizing unconscious tasks and codifying know-how, it is highly significant to organize trial-and-error iterative work into standardized flow work.
Standardized flow work fits not only tools but also systems.
Within flow work, there are tasks executed by generative AI and tasks executed by humans. By connecting these with a system, the entire flow work becomes executable.
Transforming into flow work and systemization yields several significant effects.
One is that generative AI is specialized for individual tasks, making it clear how to optimize the efficiency and accuracy of generative AI for each task.
Second, multiple workers can add knowledge to the generative AI, and the benefits extend to everyone.
Third, it becomes easy to gradually shift the division of tasks within this work to generative AI.
In this way, by transforming iterative work into flow work and accumulating the knowledge generative AI needs for each task as a system, intellectual work approaches automation like a factory line.
And by incorporating the improvements in generative AI's fundamental performance that evolve with the times, and leveraging the accumulated knowledge specialized for various tasks, it will become possible to make the entire flow work an automated process driven by generative AI.
Virtual Intelligence
This concludes the analysis from the perspective of iterative work and tools, and flow work and systems.
Another article I recently wrote further develops this discussion.
In that article, I touched upon the topic of orchestration by virtual intelligence.
Currently, and in the very near future, due to performance limitations, generative AI performs better in terms of efficiency and accuracy when focused on specific tasks.
Therefore, as discussed earlier with flow work and systems, an ideal mechanism was to connect specialized generative AIs for each individual task through a system.
However, even if generative AI's performance significantly improves, it might be more efficient and accurate to process by switching roles and utilizing different knowledge within a single processing run, rather than simply processing various tasks in parallel.
This method would eliminate the need for a system to link generative AIs together. Operations similar to system integration would occur within the generative AI itself.
Furthermore, from a situation where task rearrangement or additions are impossible without system changes, the generative AI itself will be able to respond flexibly.
This means returning the flow-worked and systematized tasks back to iterative work.
However, the iterative work that returns after undergoing this flow-working and systematization process will be in a state where reusable knowledge has been formed, even if the number of generative AIs is increased or their versions are changed.
This resolves the problems of human iterative work and enables the performance of flexible tasks similar to those done by humans.
Here, I call the ability of generative AI to switch roles and knowledge during a single execution "virtual intelligence." This is analogous to a computer's virtual machine.
Just as virtual machine technology simulates entirely different computers running on a single piece of hardware, a single generative AI processes by switching between multiple roles.
Generative AI has already naturally acquired this virtual intelligence capability. This is why generative AI can simulate discussions involving multiple people or generate novels featuring multiple characters.
If this virtual intelligence capability improves and is provided with sufficient knowledge, it will become possible to perform iterative work.
Intelligence Orchestration
Furthermore, I refer to the ability to freely combine multiple roles and knowledge to perform tasks as "intelligence orchestration."
This is similar to orchestration technology that handles multiple virtual machines.
Just as orchestration technology efficiently operates systems by launching necessary virtual machines when needed, a generative AI with improved intelligence orchestration skills—a capability of virtual intelligence—will be able to flexibly perform iterative work, maintaining efficiency and accuracy while appropriately utilizing numerous roles and knowledge.
Symphonic Intelligence
Generative AI that reaches this stage can be called Symphonic Intelligence.
Just as an orchestra, with each musician proficient in their instrument, plays a single piece while fulfilling their respective roles, Symphonic Intelligence can play a symphony of intellectual work.
This Symphonic Intelligence is a new concept, representing an endpoint for generative AI.
However, Symphonic Intelligence itself already exists.
It is our human intelligence.
It is precisely because we possess Symphonic Intelligence that we can unconsciously perform complex intellectual tasks flexibly through iterative work, utilizing a multitude of know-how.
Finally: The Form of AGI
By providing generative AI, capable of simulating Symphonic Intelligence, with flow work processes and knowledge bases for other tasks, it will be able to handle multiple iterative tasks.
As it becomes capable of handling a multitude of different iterative tasks, it will likely grasp common principles and structural patterns in knowledge across those tasks.
At that point, for entirely unknown iterative tasks, with just a simple explanation from a human, it will be able to learn the know-how of that task simply by observing how a human performs it.
This is true Symphonic Intelligence. Once this stage is reached, humans will no longer need to expend effort on flow-working or codifying know-how.
Furthermore, the knowledge thus automatically accumulated by generative AI can be shared among generative AIs themselves.
When that happens, generative AI's learning capacity will far surpass that of humans.
This can be said to be one form of AGI.