No XAI without XI, no XI without I²: ethical paradoxes and borderline situations in Artificial General Intelligence

The aim of this position paper is to explore the ethical boundaries of Artificial Intelligence (AI), considering both its imminent yet unrealized developments. By highlighting three paradoxes and three borderline scenarios featuring Artificial General Intelligence (AGI), our topic is intended to be enriched with futurology methods such as Horizon scanning and Futures Wheel. While data is considered foundational in every scientific field, an imaginative effort is recommended in this particular case to stay ahead of what awaits us and to develop more effective, proactive strategies. Although no solutions are proposed, a scenario has been presented where space-partitioning (whether at the level of towns, countries, or continents) would simultaneously preserve the personal freedom of choice for citizens regarding whether to live in a society pervasively influenced by AGI. This approach would also enable the identification, isolation, and timely resolution of potential degenerations and issues foreshadowed by the aforementioned paradoxes.