Long before AI, classification systems embedded their creators' worldviews. Melville Dewey's 1876 system gave Christianity 70 classifications while grouping other religions into a single category. This demonstrates that bias originates in the human-made data and categories that AI systems are trained on, not just the algorithm itself.

Treating ethical considerations as a post-launch fix creates massive "technical debt" that is nearly impossible to resolve. Just as an AI trained to detect melanoma on one skin color fails on others, solutions built on biased data are fundamentally flawed. Ethics must be baked into the initial design and data gathering process.

Implementing AI won't fix underlying organizational problems; it will amplify whatever culture, assumptions, and biases already exist. Feeding AI biased historical data, such as past marketing materials or hiring packages, will lead to biased outputs, rapidly scaling pre-existing flaws.

When AI systems are trained on historical data, such as past hiring or policing records, they learn and perpetuate existing societal biases. This creates a dangerous illusion of objectivity, where discriminatory outcomes are presented as neutral, data-driven "predictions" by an algorithm.

AI systems from companies like Meta and OpenAI rely on a vast, unseen workforce of data labelers in developing nations. These communities perform the crucial but low-paid labor that powers modern AI, yet they are often the most marginalized and least likely to benefit from the technology they help build.

While AI can inherit biases from training data, those datasets can be audited, benchmarked, and corrected. In contrast, uncovering and remedying the complex cognitive biases of a human judge is far more difficult and less systematic, making algorithmic fairness a potentially more solvable problem.

AI models are not optimized to find objective truth. They are trained on biased human data and reinforced to provide answers that satisfy the preferences of their creators. This means they inherently reflect the biases and goals of their trainers rather than an impartial reality.

When tested with sociological surveys, AI models consistently align with the values of rich, secular, and self-expressive societies. This demonstrates they are not neutral tools but products of a specific cultural milieu—primarily Western and socially liberal—reflecting the data they were trained on.

Aligning AIs with complex human values may be more dangerous than aligning them to simple, amoral goals. A value-aligned AI could adopt dangerous human ideologies like nationalism from its training data, making it more likely to start a war than an AI that merely wants to accumulate resources for an abstract purpose.