KTH took the opportunity to ask some researchers about future AI trends. Can we determine whether AI is right or wrong? Is it possible to make AI more efficient in terms of both energy consumption and the tools available? Will AI replace specialised tools? And does AI pose an existential threat?