Many of us are scrambling to leverage GenAI, but we are often not aware of the many risks that may arise. As we undergo this journey, it would be helpful to have a guide. To that end, I would like to see if we as a community can write the following book:
As we move towards birthing our new AI baby, it will be important that we parents understand how to properly govern and mitigate the risks associated with our new GenAI baby. Let's understand how we can raise our GenAI babies in a healthy, safe, and secure manner.
Month 1: How do I determine what "age of development" my AI baby is? This helps set expectations and will answer questions around why size matters and how various frameworks can triangulate the AI baby's age.
Month 2: How can I calibrate my expectations around the true capabilities and limitations of my AI baby? We often mistake performance for competence. Given the age and how it is trained, this will help us understand the bounds and limits of the AI baby.
Month 3: What are indicators of concern as my AI baby gets more incorporated into my business? We need to be prepared for how our businesses may be impacted by the unexpected actions and behaviors of our AI baby.
Month 4: When might I need to call in specialists? Once born, our AI babies are going to promptly get sick with some virus or injection. We need insights on which illnesses various AI security vendors might address.
Month 5: How do I make sure that my AI baby stays within the bounds of my family/business? One of the greatest influences on the AI baby will be its friends through retrieval augmented generation and vector databases. Let's make sure it associates with the right friends.
Month 6: How do I diagnose what's normal from what's novel with my AI baby? Some emerging needs can be easily anticipated and planned for. Others might be novel and require further attention. This will help us differentiate.
Month 7: How should I enlist my AI baby in the defense of my business? Before we give a weapon to our AI baby, let's make sure we know where it can best serve in our defense.
Month 8: How should I plan for the teenage years when my AI baby will have a mind of their own? When the AI baby turns into a teen, let's make sure that it doesn't turn us all into paperclips.
Month 9: How should I select my AI babysitter? We all can use help and the AI frontier companies are there to help keep our AI baby safe.
Bonus chapter! How do I deal with my surrogate AI baby? We have vendors that sell us a product with its own AI baby. We need to ask the right questions before we take custody of an AI baby that we did not birth.
Despite all the challenges ahead, if we can learn from each other's mistakes and plan ahead, we’ll have healthy, happy AI children that will take good care of us when we are older and they take over the world.
