AI should expand agency, not hide it

AI is often discussed as though there are only two futures: total automation or total rejection. Solarpunk offers a more demanding third option. Build technology that increases people’s ability to participate in their own work and the decisions that shape their lives.

The Scopus literature synthesis highlights digital platforms and interactive media as possible tools for inclusion, transparency and the rehearsal of alternative futures. It also names the risks: digital exclusion, privacy problems and technology used as superficial branding. A good interface cannot erase those trade-offs.

So we prefer an augmentation question over an automation question. Instead of asking how to remove the researcher, teacher or learner, ask which repetitive work can be reduced while leaving interpretation, judgment and learning visible. This is the thread connecting LitSynth, InstantGrade, Root and the experimental systems around them.

That approach creates obligations for the builder. We have to describe a system’s limits honestly, preserve meaningful user control and avoid presenting a generated answer as settled truth. In research, that means provenance. In education, it means feedback a teacher can inspect. In learning, it means a person can follow the path rather than simply receive an answer.

AI can be part of a more capable future. But capability is not delivered by a model alone. It is designed into the relationship between a person, a tool and a real task.

— SolarPunkWorks Pvt Ltd