> For the complete documentation index, see [llms.txt](https://projectsolo.gitbook.io/solo/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://projectsolo.gitbook.io/solo/use-cases.md).

# Use Cases

## Primary example: AI data labeling, HITL, and RLHF

AI/data companies depend on human judgment. Label quality, preference data, safety evaluation, emotional response, cultural context, and real-world observation all become weaker when the platform cannot prove that the contributor is a real, unique human.

### Problem

A data platform may receive thousands of labels, rankings, reviews, or feedback submissions. Some may come from real users. Others may come from bots, duplicate accounts, account farms, scripts, low-effort workers, or synthetic respondents. These bad signals can look valid in the database while quietly degrading model quality.

### How Solo fits

1. **Before the task:** gate task access to verified humans only.
2. **During the task:** use continuous or random re-verification for longer sessions and high-value labeling workflows.
3. **At submission:** associate each accepted label, score, review, or feedback package with a proof-of-humanity check.
4. **After the task:** audit datasets by checking which work came from verified human sessions.

### What the customer gets

* Cleaner human feedback pipelines.
* Lower bot and duplicate-account risk.
* More auditable proof that responses came from verified humans.
* Better protection against account sharing during longer work sessions.
* A verification layer that does not require the data company to store biometric records.

## Other use case examples

* **Financial services:** proof of humanity at onboarding or high-risk transaction checkpoints as a privacy-preserving anti-fraud layer.
* **Healthcare and telemedicine:** confirm that the patient or participant is present for sensitive digital interactions.
* **Government and civic technology:** support one-human-one-claim or one-human-one-vote style services without making document identity the default proof layer.
* **Airdrops and reward campaigns:** reduce Sybil farming by enforcing one verified human before claims, points, or incentives are distributed.
* **Governance and voting:** support humanity-gated voting or one-human-one-vote modules without exposing legal identity by default.
* **Gaming:** verify human presence at login and at reward milestones in session-based products.
* **DePIN contribution:** verify contributors before accepting physical-world data submissions.
* **SocialFi and reputation:** reduce fake accounts and support private, portable reputation systems.