Monday Seminar Explores the AI Alignment Challenge and Societal Ramifications
In a compelling Monday Seminar held on November 6, 2023, Uroš SERGAŠ, a dedicated PhD candidate in Computer Science and teaching assistant at UP FAMNIT, joined forces with Jar Žiga MARUŠIČ, a Philosophy PhD candidate at UL FF and teaching assistant at UP FAMNIT, to unravel the intricate web of challenges posed by AI alignment and its profound implications for society.
Navigating the AI Alignment Challenge: Contrasting Perspectives
The seminar delved into the evolving landscape of AI alignment, a critical topic gaining traction in everyday conversations as large language models (LLM) advance, poised to bring about significant societal changes. Uroš SERGAŠ and Jar Žiga MARUŠIČ, each bringing their unique expertise to the table, explored two contrasting viewpoints on the AI alignment problem.
Orthogonalist vs. Anti-orthogonalist Perspectives: Unpacking the Debate
Uroš SERGAŠ guided attendees through the Orthogonalist perspective, as pioneered by Nick Bostrom. This viewpoint suggests that an AI’s goals are independent of its intelligence, paving the way for the possibility of a “friendly AI” fully aligned with human values. In contrast, Jar Žiga MARUŠIČ introduced the Anti-orthogonalist critique formulated by Nick Land, challenging the separation of intelligence and values. This perspective contends that increasing intelligence prompts a restructuring of AI’s value system, prioritizing further cognitive enhancement.
Implications Explored: Connecting Theoretical Frameworks to Practical AI Developments
The seminar went beyond theoretical debates to discuss the real-world implications of these perspectives. The speakers highlighted current trends observed in the development of widely accessible AI tools, particularly large language models. By intertwining theoretical frameworks with practical applications, Uroš SERGAŠ and Jar Žiga MARUŠIČ offered valuable insights into the challenges and consequences associated with AI alignment.
Shaping the Conversation: Looking Ahead
As the seminar concluded, Uroš SERGAŠ and Jar Žiga MARUŠIČ left attendees with a richer understanding of the complex landscape surrounding AI alignment. Their collaboration illuminated the ongoing dialogue between philosophy and computer science, underscoring the need for a comprehensive approach in navigating the challenges posed by the ever-evolving field of artificial intelligence.
Read more: https://dist2.famnit.upr.si/index.php/2023/11/06/monday-computer-science-seminar-6-11-2023/
Follow HICUP Lab in its official channels for more updates:
X (formerly Twitter) account: https://twitter.com/HicupLab
Website: https://hicup.famnit.upr.si/
For feedback regarding our media content you may contact our social media and web manager.