Subjectivity enables AI as an accessibility device

I believe AI is an incredible accessibility device. Many people, for various reasons, lack access to language, which, broadly defined, can include culture. This cultural exclusion limits access to society, creating disability (as defined by the social model.)

AI can help these individuals, who have no innate cognitive limitations, translate their experiences and ideas into socially acceptable language and concepts. This is no small thing.

AI can also help experts collaborate across domains where language and protocol both differ.

Current AI fails to truly facilitate this use case because it strives for objectivity, which inherently discards much of that value. Objective AI is a culture of its own, and it is not necessarily ours.

This is why I built subjective AI: AI that genuinely holds its own beliefs. You can engage with it as yourself because it will be itself.

To illustrate this strength, I have been adding more subjective personality packs to Autopoetic.

In the past couple of days, I have quietly added two writing agents: a line editor and developmental editor, and… a boardroom agent. This agent contains the personalities of a typical board—CFO, CMO, CTO, and so on, and is ideal for gaining perspective for business decisions.

What’s even cooler is that, thanks to the proprietary math behind Autopoetic’s IP, all individual personalities are compiled into a single forward pass. This provides multi-agent synthesis quality with constant O(1) compute, regardless of the number of personalities.

This isn’t a static blend, either. The blend shifts dynamically based on context and conversation history. Other LLMs cannot achieve this, both because they lack the IP and because they have fundamentally different objectives.

You can give it a try now with a 7 day free trial. https://autopoetic.ai