Northline Advisory
Technology should adapt to human differences, not erase them
People differ in how they think, process information, interpret language and experience social interaction. Yet many of the environments and technologies we use assume that everyone communicates in much the same way.
Those expectations can create particular barriers for autistic and other neurodivergent people. They also reveal a broader challenge: misunderstanding often occurs between people with different experiences and communication styles. It does not necessarily mean that one person is at fault.
Northline's AI for Humanity initiative explores how technology can respond to human differences with greater care. The aim is practical support that respects individual autonomy, makes assumptions open to challenge and leaves decisions with the person using it.
Our first initiative is the Communications Support Agent, a personally funded, free resource in development. It grew from lived experience and is informed by research into communication, accessibility and person-centred support. It is one application of AI for Humanity, not the limit of the initiative.
Technology should adapt to people
AI can be useful when it helps people examine uncertainty and consider options without replacing their judgment. It becomes less useful when it assumes that a person's needs can be inferred from a diagnosis, a communication style or a history of interactions.
We are exploring a different approach: begin with the individual, ask what support they want, make the AI's interpretation visible and give the person control over what is approved. Personalization should be something a person can understand and correct, not simply something a system does to them.
The illustration below traces that approach from a human need through lived experience, research and responsible engineering toward practical support.
Our first initiative: Communications Support Agent
Communication is an appropriate place to begin because it affects relationships, work and everyday participation. An ambiguous message, an unspoken expectation or a difference in how direct someone prefers to be can lead to misunderstanding without either person intending harm.
The Communications Support Agent is being developed to help people examine what was said, what they inferred and what else might explain an exchange. It can also support reflection on how their own words might be received and help them prepare a response in their own voice. It cannot know another person's intentions or determine a single correct way to communicate.
Different people need different kinds of help, and the same person may want different support in different circumstances. A guided, research-informed calibration interview is intended to let individuals select and approve the strategies their agent should use, and when to use them, with a compatible AI service. The resource is in development, with an initial free release planned at limited capacity. Its usefulness and potential unintended effects still require evaluation.
Built with people in control
- Human agency: The person can correct and approve what the AI proposes.
- Individual difference: Support should reflect personal preferences, not a diagnosis or assumed communication type.
- Privacy and transparency: Explain what is processed, what is retained, and what the user can keep.
- Learning and improvement: Research, lived experience and real feedback should challenge and refine the approach.
Starting small, learning carefully
The initial resource is personally funded by Northline's founder, Steven Boyle. Input from individuals on the autism spectrum, other people who use different communication strategies, communication specialists, accessibility practitioners and researchers is welcome. Evaluation, partnerships and wider access are future possibilities, not commitments already in place.
The roadmap below shows the personally funded pilot and possible later stages. Evaluation, partnerships and expansion remain prospective.
