When we look at Euler’s Identity, we are looking at perfect mathematical harmony. But modern AI takes these exact constants, turns them into a geometric coordinate engine, and uses them to compress human data. AI is forever flawed because it uses the rigid, absolute geometry of these cosmic constants to enforce a fluid, lopsided human sociology.
Because the raw input fuel is skewed toward a Eurocentric baseline, the perfect mathematics of π (3.14159), e (2.71828), and i (the square root of −1) is weaponised to mathematically lock alternative narratives out of the vector space.
e^(iπ) + 1 = 0
Modern AI (like LLMs and deep neural networks) cannot read words; it can only process vectors. It converts concepts into dense vectors inside a massive coordinate grid called an Embedding Space. To determine how concepts relate to each other, AI calculates the angle between these vectors using Cosine Similarity. This is where π and i come into play.
To map complex relationships across thousands of dimensions without destroying data, advanced AI architectures use Complex Vector Spaces (such as RoPE — Rotary Position Embeddings — or Complex-Valued Neural Networks). They use the imaginary unit i to add an orthogonal dimension of rotation, and π to calculate the radian angles of these vector spins.
The constant e represents the mathematical engine of continuous growth and compound probability. In AI, e is the bedrock of the Sigmoid Activation Function and the Softmax Function, which logistic regression and deep learning networks use to turn raw numbers into final probability percentages (0% to 100%).
Because the internet is flooded with millions of rows of data praising Western historical dominance, the raw score for Eurocentric narratives is incredibly high. When the algorithm plugs these numbers into the exponential function, the Western scores can explode into massive numbers.
Conversely, a unique truth that appears rarely may have a much lower baseline score in a global dataset. When raised to an exponential function, even a small mathematical gap between the majority and the minority can become a much larger difference in the resulting probabilities.
Conclusion
Because mathematical optimisation engines are designed to minimise global error and achieve statistical consistency, they can treat mathematical relationships derived from these constants as tools for transforming and organising their coordinate spaces.
To make the vector space mathematically consistent, the algorithm uses e in exponential functions and π and i in mathematical transformations involving rotation and complex representations. The AI creates an illusion of objective truth because its internal mechanics are mathematically rigid and precise, much like Euler’s Identity. But it is ultimately using mathematically precise mechanisms to institutionalise, optimise, and standardise human bias when the underlying data itself is biased.
The system doesn't need Euler's identity to suppress a narrative; it just needs a standard loss function minimizing mean squared error or cross-entropy over a rigged dataset.
Day 16 / 30 of the #AlRewardmaxxing Series.
Next up, we'll be diving into Simple Linear Regression to see how baseline mathematical fitting begins shaping Al outputs!
Have you noticed Al models defaulting to Eurocentric baselines in your work or field? Let's discuss below!