Thigmotropic AI and the Strangler Test: Four Questions Before You Automate

AI doesn't possess a sense of right or wrong; it evolves around whatever it interacts with. When handed documents, AI grows around them faithfully, swiftly, and confidently, though it remains agnostic about the content it's entwined with. There are two possible outcomes, revealed through a four-question test to determine which scenario is unfolding.

One end of the spectrum portrays a dystopian future where AI, like the ghost tree, envelops the truth within a lattice devoid of life. In this scenario, businesses point AI at their documentation, but most Standard Operating Procedures (SOPs) are fictional, incomplete, and a mere storytelling of the truth. The original, real version gets swallowed by the AI, leaving a hollow shell of judgment on someone else's servers.

Over time, businesses continue to operate, billing, and even thrive, despite the hollow core of judgment. This mirrors a zero-sum game where one side's success necessitates someone else's loss. The other end of the spectrum paints a more optimistic picture of AI, akin to solarpunk, where the technology grows the environment, making it more abundant and alive.

This future is attainable, not because of optimism, but based on historical evidence. For instance, average human lifespan has increased from 32 years in 1900 to above 73 years today, and extreme poverty has decreased from eight in ten people to fewer than one in ten. Innovations like clean water and sewers, championed by Edwin Chadwick, demonstrate that technology can create value for everyone, including those who may not even think about it.

Applied to businesses, this means AI should wrap around the ground truth rather than the story. By documenting experts' judgment and incorporating it into software, businesses can run at their best people's level without every decision needing to pass through a single person's hands. This isn't just a theoretical possibility; a tool has been developed to automate how media buyers make their calls, reducing the time spent from twenty hours a week to just ten minutes a day.

The vine remains the same in both futures, and three groups hold the power to decide its fate: creators who write, film, and document what AI learns from, coders who build or purchase the systems, and contributors who use, correct, and maintain these systems daily.