In Co-Intelligence (book) Ethan Mollick describes four rules that can help us work better with AI:
- Always Invite AI to the Table.
- Be the Human in the Loop.
- Treat AI Like a Person (But Tell It What Kind of Person It Is).
- Assume This Is the Worst AI You Will Ever Use.
1. Always Invite AI to the Table.
Always use AI in every task possible, regardless of whether it’s helpful or not. This constant interaction allows us to familiarize ourselves with AI’s capabilities and limitations, as AI lacks a definitive instruction manual 😊.
2. Be the Human in the Loop.
Even as AI becomes more capable over time, we have to remain active participants in the process. This includes checking for inaccuracies, providing oversight, and ensuring ethical considerations are addressed.
3. Treat AI Like a Person (But Tell It What Kind of Person It Is).
We should interact with AI as we would with any other person whilst defining a specific persona for the AI, outlining its role and the problems it should address. Providing context and constraints through persona creation helps guide the AI towards desired outputs, similar to how wording requests specifically to a person is more likely to yield better results.
4. Assume This Is the Worst AI You Will Ever Use. 🙅♂️
LLMS are improving faster than Moore’s Law (5-14 months compared to 24 months). The continuous and rapid evolution of AI capabilities will make the current version of the AI obsolete. This rapid advancement suggests that the AI tools available at any given time will likely be surpassed by more powerful versions in the future. We should approach AI with the understanding that its capabilities will continually improve.
source::Co-Intelligence (book)