Part 14/16:
Applied AI Versus Fundamental Research
He stresses that applied AI has been steadily developing over the last 60 years, but fundamental research remains essential for breakthroughs toward AGI. Industry-driven progress often neglects scientific exploration, which is critical for understanding and safely implementing more autonomous and trustworthy systems.
The Future of AI Safety and Society
Thoren warns against the dangers of black-box AI systems, such as large language models, which lack explainability and thus pose significant safety risks. Ensuring transparency and controllability is his primary concern—without it, powerful AI systems could become unpredictable, undermining societal trust and safety.