Module 4 / Lesson 12.2
AI 2027 Scenario
What Is the AI 2027 Scenario?
Learning: Learn about the AI 2027 scenario forecast. There is no single way to explore the AI 2027 scenario; this is a MUCH MORE OPEN ENDED learning opportunity than usual. Spend an hour (or more) checking it out through exploration of the website, interviews, and any other additional resources on the internet you find. As you explore, please do the following:
- Read the text below about AI 2027.
- Visit AI-2027.com, click through the timeline and explore the prediction of different things throughout the timeline. Spend at least 20 minutes exploring the page.
- Watch the Daniel Kokotajlo interview on Hard Fork. (Prof. Shore likes this interview a lot and recommends this one!)
- Spend at least 20 minutes exploring AI 2027 through other sources you choose. Check out the sources that interest you, which potentially include these suggestions:
- Google AI 2027 and check out essays, blog posts, youtube clips, newspaper stories, and anything else that interests you.
- If you want to explore more, there is a link below to an additional video.
Reflection: As you explore, note the branching points you find most influential and intriguing. For each, think of
- (a) what would have to happen for this branch to take place and what would be the impact of this happening, and
- (b) what would an alternative branch look like.
Imagine waking up in 2027 to find that artificial intelligence has become superhuman across many cognitive domains. Governments, firms, and societies would face immediate, once-in-history choices, choices that could lead to unprecedented benefits or to irreversible catastrophe.
This is the core of the AI 2027 Scenario: a detailed thought experiment (interactive timeline at AI-2027.com) created by researchers and futurists to explore what happens if AI capabilities accelerate far faster than expected. It is a stress test, not a prediction: it asks “what if” to surface assumptions, risks, and decision points.
The idea has traction among advocates who call for urgent safety measures (e.g., PauseAI) and among thinkers such as Stuart Russell and Eliezer Yudkowsky who emphasize the plausibility of rapid, high-stakes acceleration. Use the scenario to practice reasoning under deep uncertainty, not to adopt it wholesale.
How to Navigate the Scenario
The official site presents a month-by-month, interactive timeline with branching points. places where different choices or small events push the world down very different paths. To get the most out of it, you should treat it as if you are walking through a branching future:
-
Start with the Timeline Overview
When you first open the site, skim the full timeline to see how it is structured. Notice the different “forks in the road” where events diverge depending on key decisions (for example: governments regulating AI vs. letting the market race ahead). Think of this like a choose-your-own-adventure book: you don’t just follow one path, you examine several. -
Focus on the Turning Points, Not Every Detail
It’s easy to get lost in small narrative elements. Instead, concentrate on the critical branching points, moments where a single decision has cascading consequences. These are usually framed as policy decisions (pause vs. race), technical events (breakthrough vs. bottleneck), or social dynamics (cooperation vs. conflict). These are the “pressure points” the scenario is trying to reveal. -
Pay Attention to Assumptions
At each branch, ask: What is this scenario assuming? For example, if one branch says “labs release an open-source AGI,” the assumption is that companies won’t prioritize safety or secrecy. If another branch says “international agreement slows progress,” the assumption is that nations can actually coordinate despite competition. Identifying these assumptions is the key skill this scenario is teaching. -
Compare Divergent Outcomes
Don’t just follow one path from beginning to end. Click back, take a different choice, and see how the downstream consequences differ. This exercise shows how small shifts early on can magnify into radically different futures by 2027 , either safer, slower development, or uncontrolled runaway growth. -
Look for Themes Across Branches
As you explore multiple outcomes, try to notice repeating themes
This helps you see how the scenario frames risks, and also where it might be oversimplifying.