Creating and exploring AGI scenarios

It's time to retire the paperclip maximiser. That thought experiment for the evolution of AGI (Artificial General Intelligence) has attracted lots of attention over the years, but it's too easy for people to dismiss it as abstract, unrealistic, or science fiction.

It's the same with many other popular narratives about the rise of AGI. These narratives often do more harm than good, hindering people from gaining a genuine understanding of the rich spectrum of possibilities ahead.

This online workshop is a chance to leave these narratives behind. It's a chance to draw on the best of recent thinking around the world, in order to jointly create and explore better scenarios for the evolution of AGI. The goal is to develop scenarios which have a good chance of striking reviewers as credible, real-world, engaging, and worthy of further investigation.

One or more of these scenarios will explore how events can, credibly, slide from an initially positive state (with optimistic participants) into a disaster.

Additionally, one or more scenarios will explore how events can, credibly, progress from an initially confused state into a successful governance of the transition to extraordinarily beneficial AGI.

== Taking place on Remo with Telepresent ==

As an experiment, this online event is taking place on Remo with the kind support of Telepresent.

On this occasion, as a trial of the system, the event will be free of charge.

Please visit futurists.telepresent.online to register for this event.

To take part in the event, you will need a desktop or laptop computer, running Chrome, Safari (on Mac only), or Microsoft Edge (on Windows only). The above page will give you the chance to check your system's suitability before the date of the event.

== More advice on using the Telepresent system ==

See this one-page PDF for more advice on having a good user experience with the Telepresent system.

== Schedule ==

(All the following times are UK times)

15:40 - System opens, ready for informal online networking

16:00 - Scene-setting presentations and general questions

16:25 - Audience split into groups, to develop their own ideas about the evolution of AGI

16:50 - Groups report their findings to the plenary

17:15 - General discussion

17:30 - Ongoing informal networking, for those who wish to stay

== Background review material ==

The following will be useful for participants to review ahead of the event:

• An example scenario: "Artificial Escalation" - a video by the Future of Life Institute
• "An Overview of Catastrophic AI Risks" - a 54 page analysis by Dan Hendrycks, Mantas Mazeika, and Thomas Woodside of the Center for AI Safety
• Survey results for round 1 of "Key open questions about the transition to AGI"
• "Global cooperation urgently needed to govern risks of advanced AI, warns new report" - press release from The Millennium Project
• A couple of activist sites: PauseAI and StopAGI
• A summary of areas of agreement and disagreement among people actually creating AGI: "Towards best practices in AGI safety and governance" (38 page PDF)