TDF5

Humanity more or less unexpectedly stumbled into a future that no one would have considered remotely realistic before: Models that learn language structures have evolved into thinking machines.

Key players consider it possible and likely that these algorithms will possess abilities superior to those of humans. The debate centers more on when this will happen than on whether it will: a matter of months or decades. It is therefore high time to prepare for it.

The visions of the future could not be more different.

• Optimists predict nothing less than the end to all scarcity. The "last invention humanity will ever make itself" will catapult us onto a new path of growth: Scientific discoveries that would otherwise take decades of human research could be realized in just a few years. The Promises: AI models could provide us with an abundance of energy e.g. through fusion reactors and hydrogen production, and drastically extend our lives through advances in medicine. The ability to automate human activities is gradually leading us—through the replacement of information-based work and the development of robotics—into a world free of labor and coercion.
• Pessimists point above all to the insane energy demands that are exacerbating the climate crisis. The displacement of labor will result in struggles over redistribution and ultimately could lead to a collapse of the market. The prospect of weapon systems with superhuman capabilities and new strategic programs are already increasing the risk of war, as the bloc that is the first to acquire a certain level of AI-capacities threatens to become invincible. New technological advancements are leading to total surveillance, and AI applications trained on human psychology and neurology are being used for behavioral control and crowd management. Unpredictable disasters loom due to the fundamental uncontrollability of these systems and the impossibility of programming them to stable follow ethical principles.

In these dynamic times, predictions about the future are particularly uncertain and it is highly likely that expectations and extrapolations will be very wrong. Nonetheless society has to decide and act now. After the presentation there will be hopefully time for discussion. Can and if so: how should AI revolution be regulated or even slowed down and what opportunities are there? Are there methods safeguarding the inherent risks? How can we avoid that AI will become a tool for or masking of dominion? How can we avoid that AI-algorithms become private property of a few monopolists that will own the world?