StillYours

What is this forecast app?

A retained-skill challenge round

It runs short, paid and human-reviewed skill challenges for work usually done with machine assistance, then separates safe normal support from a genuine capability gap and funds practice instead of punishment.

• The worker chooses a bounded capability and reviews the support that will be available and temporarily withheld.
• A qualified human prepares a safe paid task with one unfamiliar variation and clear stop conditions.
• The worker completes the task, explains choices and identifies when assistance would be unsafe or insufficient.
• The reviewer records only the demonstrated capability, needed support and next paid practice date.

The result: The worker and accountable operator know which skill is retained, which support is needed and what paid practice must happen next.

Why it is on the list

Assistance can conceal skill decay

By 2031, people may produce impressive work with dependable assistance while getting fewer chances to notice an edge case or recover when the help disappears. Essential services still need to know where human agency is real. StillYours makes that question answerable without pretending all support is cheating or giving a machine power to end a career.

Why 2031—not 2026?

Competency tests exist today. This forecast assumes assistance is so normal that workplaces need recurring, worker-controlled checks of the boundary between supported performance and retained human recovery skill, with paid remediation and accessibility protection built in.

Why people would return: Retention decays and task boundaries change, so evidence must be refreshed at intervals based on actual risk.

What would have to change in the world?

What makes it more than better AI?

The service governs human agency, support boundaries and access to physical practice; model performance alone cannot answer those questions.

Conditions that must exist:

• Assistance becomes normal enough to reduce unaided repetitions.
• Rare edge cases still require human judgement.
• Non-punitive paid review and remediation are enforceable.

When this forecast fails: If assistance remains reliably available across all relevant failures or ordinary work supplies enough unaided practice, recurring challenges are unnecessary.

How it could be built

The service, technology and institutions it would require

Create worker-controlled capability intervals, safe task variants, human observation, accessibility protection and automatic paid remediation.

Essential dependencies

The hardest part: Testing genuine retained capability fairly when legitimate assistance and accessibility tools are part of how the worker actually performs.

A simpler alternative: Regular paid refresher training run by the employer.

Risks and limits

What could go wrong?

When it must stop: The worker or assessor can halt immediately; no adverse result is recorded and pay continues.

Why this position

Why StillYours is ranked #7

Its 2031 logic, repeat use and callable finish are strong. It ranks seventh because the safety problem is unusually sharp: a badly designed challenge could become ableist, punitive or falsely precise, and different places may disagree about which unaided skills must actually be retained.

Why it outranks the next forecast: It ranks above ProofFerry because it directly creates fresh evidence through observed work and includes paid remediation; ProofFerry mainly governs whether other organisations will accept evidence created elsewhere.

Strongest counter-case: The service could revive exclusionary testing and ableism in the language of resilience. Designing systems that fail safely may be better than demanding that workers preserve rarely used unaided abilities.

Rank range across tested weights: 2–9. The exact rank is an authored judgement, not a measured probability.

Evidence behind the forecast

Current sources and their limits

Observed and published evidence grounds the world pressures and present constraints. The category, product, developer, reviews, rating and exact rank are fictional forecasts and may be wrong.

Browse the complete source register →