What implications does the possibility of hacking reward systems have for failure modes of AI, humans, and civilisation as a whole?

Join us as Anders Sandberg talks about wireheading. Will be of interest to anyone interested in AI, philosophy, history of ideas, big-picture thinking about civilisation.

Register for free at https://www.eventbrite.com/.../talk-wireheading-risks...

**Talk Abstract**
The classic "wireheading" experiment took place in 1952 when James Olds stimulated the septal area of the rat brain, finding that the rat would press a lever again and again in order to get a stimulation. Not long after the experiments, people started worrying that, with advances in neuroscience, humans might modify their own brains to trigger its reward systems on demand becoming wireheads. This lead to fear that this would be a threat to civilization itself, an existential threat. This talk will explore the wider history , philosophy, and myths of wireheading, showing how it links up to concerns about the stability of humans and civilization, and becomes a practical problem in modern machine learning.

**Speaker Bio**
Dr Anders Sandberg is a Senior Research Fellow at the Future of Humanity Institute and the Oxford Martin School, and Ethics and Values Fellow at Reuben College, University of Oxford. Dr Sandberg's wide-ranging research centres on estimating the capabilities and underlying science of future technologies, methods of reasoning about long-term futures, existential and global catastrophic risk, the search for extraterrestrial intelligence (SETI), as well as societal and ethical issues surrounding human enhancement.