Confronting Impossible Futures
๐ Abstract
The article discusses the lack of strategic planning by organizations for the future of AI, despite the rapid advancements in AI capabilities. It explores the reasons why organizations are not seriously accounting for the possibility of continued AI improvement in their long-term planning, and provides suggestions on how to better prepare for potential AI futures.
๐ Q&A
[01] The Lack of Strategic Planning for AI's Future
1. What are the key reasons why organizations are not seriously planning for the future of AI?
- There is a lot of uncertainty about the future of AI, with some believing in exponential growth in capability while others think AI has reached its limits.
- Many discussions about the future of AI focus on outcomes that seem far-fetched or outright marketing hype, like the possibility of superintelligent machines, which makes it difficult to plan for.
- There is a lack of clear documentation and understanding of the current capabilities of AI systems, leading to underestimation of their abilities.
- The jagged nature of AI capabilities, where it excels at some tasks but struggles with others, makes it easy to focus on its limitations rather than its impressive achievements.
2. What are the potential futures for AI that organizations should be planning for? The article outlines four potential futures for AI:
- A capabilities plateau
- Linear growth in capabilities
- Exponential growth in capabilities
- Achieving Artificial General Intelligence (AGI)
The author argues that all of these possibilities remain in play, and organizations need to plan for multiple contingencies rather than just picking one official future.
[02] Strategies for Preparing for AI's Future
1. What techniques does the article suggest for organizations to explore different AI futures? The article recommends using scenario planning, a strategy research tool that allows organizations to examine how their strategies might play out in different future worlds. This can be done at the organizational or even personal level.
2. How can AI, specifically GPT-4, be used to assist with scenario planning? The article states that GPT-4 can do the "heavy lifting" of scenario planning, providing meaningful insights about how an organization's strategies may play out in different future scenarios. This can be a more efficient alternative to the typically time-consuming process of learning and conducting scenario planning exercises.
3. What is the overall message the article conveys about preparing for AI's future? The article emphasizes the need for organizations to stop "pretending that the world isn't changing" and to start taking control to steer towards a better future. While the exact future is uncertain, the article encourages proactive planning and adaptation to the rapidly evolving AI landscape.