Smart Machines and Systems at the Service of Mankind: Timo Seppälä - “Contemporary digitalization: AI diffusion monitoring among S&P500 companies”

PhD Timo Seppälä is a University Lecturer (Operations Management) at Aalto University. He earned his PhD in industrial economics and management. Among other things, his scholarly work has touched upon global value chains, digital platforms, and lately the semiconductor industry. He serves as a chairman of the board at Koja Yhtiöt Oy & Koja Oy and member of the board at Koja USA Inc. and Chiller Oy (An Aalto Family Company).

The topic discussed during this conference was “Contemporary digitalization: AI diffusion monitoring among S&P500 companies”.

When it comes to technology diffusing, the keynote speaker explained, it takes a rather long period of time for a company to actually implement a new technology to their products or services. There would be several phases before that, e.g., a period of learning-by-observing and discussions of how these new technologies could be implemented. If one takes ChatGPT, which was launched roughly two years ago, as an example, then it could be said that it would still take approximately 4-6 years to understand what is the exact value of this system for companies as such. This understanding, of how much time it would take from discussing to implementing, was reached after analysing 300 000 S&P500 companies management speeches and news items from the Reuters database.

Moving deeper into the topic, the keynote speaker introduced the timeline of development from connected products to computerised products - now, around 2014-2015, the products were implemented with a connectivity and became smarter. After connectivity implemented the companies were able to start collecting data. Mr Seppälä brought in John Deere tractors as an example.

Taking steps from connected and computerised to autonomous products, in 2019, after some time of data collecting, John Deere came forward with the idea of a “driverless” tractor.  In just three years time, they revealed a fully autonomous tractor. Farmers in Finland, the speaker stated, who operate these tractors, have acknowledged the benefits of them. It has allowed the farmers to concentrate on other - more meaningful - tasks, while not having to drive the tractor themselves. Moreover, John Deere, has stated that they wish to continue their transformation towards autonomous farming. What this means is that the tractor will take more tasks from the farmer and start, for example, ordering materials or planning the supply chain. Perhaps even production and delivery planning. In this way, the speaker explained, the tractors will start replacing the other software systems within the companies by taking their roles. This will lead to enterprise software being replaced by the software in these smart, autonomous products.

In the last decade, while implementing AI, machine learning, machine vision, and large language models to our products, we have been able to reach a point where tractors, mining equipment, etc., can be operated without a human in the driver’s seat. Mr Seppälä added that for industrial companies it has been easier to implement these technologies, due to the use cases of this machinery being much simpler than that of autonomous vehicles.

Another question that has been posed is when do these autonomous products become AI-Agents. Simply put, once machine learning, machine vision or large language models have been introduced into the technology stack of a computerised product - it can be called an AI-Agent.

What is next in product autonomy? What should be done to have these technologies more widely available in the coming years? One opportunity would be to isolate the autonomous vehicles from the current system of traffic. To create for them a system inside the system. Or instead remove the human quotient from the equation so that the whole traffic system could become autonomous? The questions are many and all of them have a myriad of possible answers.

You can find full discussion on our Youtube channel. More information at conference homepage