Technology helps us move forward. We choose the course.
At Progymédia, we have been developing software for decades, and artificial intelligence now plays a central role in our work. We use it every day, we build systems that use it, and we strongly believe in its potential. In fact, this article was written with assistance from ChatGPT.
But despite my enthusiasm, I think it is time to ask ourselves a fundamental question: how far are we willing to let artificial intelligence go?
Automation is part of our evolution
For thousands of years, people have invented tools to accomplish tasks more efficiently.
The Industrial Revolution transformed our capacity to produce goods. Computers automated information processing, and the internet then reshaped how we communicate and collaborate.
Artificial intelligence continues that progression. It now allows us to automate tasks that previously required human intellectual abilities.
That is an extraordinary advance, and I believe it can contribute a great deal to our society.
But technological progress does not necessarily mean social progress.
I am concerned about the economic costs of AI, its environmental impact and, above all, the social upheaval ahead if we move forward without thinking about the consequences.
We need to ask what place we want people to have in a world where machines can increasingly act on their own.
A boundary: agency
Agency is a person’s ability to make decisions independently, exercise judgement and act intentionally to influence their environment and achieve their goals.
The concept has been around for a long time, but with the arrival of AI, it takes on a new dimension.
We are now developing agents that can analyse situations, plan, make operational decisions and carry out complex tasks without human intervention at every step.
And I think that is an excellent thing.
But there is a fundamental difference between giving AI the autonomy it needs to accomplish a task and letting it determine the goals it should pursue.
We can delegate execution and operational decisions while retaining control over goals, values and their consequences.
That is precisely where I want to draw the line.
Our decision at Progymédia
We will continue to develop increasingly capable and autonomous AI agents. We will continue to automate what can reasonably be automated.
But I have decided that our guiding principle will be clear:
AI can have autonomy in execution. Agency, however, must remain human.
That means our systems must continue to serve goals defined by people, within a framework they can understand, control and question.
I know that this boundary will not always be easy to define. I do not claim to have all the answers or to know what the future holds.
But I know the direction I want to take.
I will do my best to ensure that artificial intelligence increases our ability to act without ever making us give up our ability to choose.
And to ensure that agency remains human.
The guiding principle I want to keep
✓ Goals and values defined by people.
✓ Autonomy in execution within a framework people can understand, control and question.
✓ A greater ability to act, with the freedom to choose.
