PublicAI

PublicAI is a decentralized AI data infrastructure platform that positions itself as "the human layer of AI." Founded in 2023 and headquartered in San Francisco, the project addresses one of machine learning's most persistent bottlenecks: the production of high-quality, human-verified training data at scale. Rather than relying on centralized data labeling services, PublicAI crowdsources this work across a global contributor network — compensating participants in $PUBLIC tokens tied directly to enterprise client revenue.

Core Mechanism

At the heart of PublicAI is a Byzantine Fault Tolerant (BFT) consensus algorithm applied to data validation. When a contributor submits a data sample, it first passes through AI-assisted automated screening. Samples that clear this initial filter enter a human voting phase where they must achieve consensus from validators at one of three thresholds: 50 votes from Scouts, 5 from Guards, or 1 from a Judge. Accepted data is marked as "Authentic" or "Helpful"; rejected data is discarded and the submitter's staked collateral may be slashed.

This two-phase validation keeps costs manageable — AI handles the bulk of obvious decisions, while humans adjudicate edge cases requiring genuine judgment. The result is a quality assurance system that is adversarially resistant without centralized review.

Contributor Hierarchy

PublicAI structures its contributor community into five tiers:

Newbies enter through referrals and complete an onboarding process before earning privileges. Scouts are proven contributors who have demonstrated quality output and form the backbone of data generation. Guards are trusted network builders who unlock referral rewards, flag malicious content, and serve as candidates for promotion to Judge. Judges are the final arbiters for disputes and complex edge cases, typically promoted from Guard ranks. Datababies are early NFT holders who serve as community anchors regardless of their active contribution level.

Progression through these tiers unlocks higher earnings, referral bonuses, and governance rights. The system also functions as quality control: low-quality actors burn through staked collateral before reaching higher tiers, while high-quality contributors accumulate economic incentive to protect their reputation.

Supported Data Types

The platform processes four primary data modalities:

• Web3 and AI training data: Content sourced from social platforms (particularly X/Twitter) used to fine-tune large language models in specialized domains
• RLHF training data: Reinforcement Learning from Human Feedback datasets, where contributors evaluate AI outputs to improve model alignment
• Text-to-Speech (TTS) data: Speech-text aligned recordings with phonetic and prosodic annotations across 44+ languages, including underrepresented languages
• Aesthetics assessment data: Human consensus judgments on the artistic quality of images, videos, and music

Solana Integration

PublicAI's most direct Solana connection is its RLHF on-chain consensus protocol. Previously, contributor voting records were stored in centralized databases. To decentralize this layer, PublicAI deployed smart contracts on Solana to record voting hashes and timestamps immutably. Contributor responses are hashed with timestamps each day, those hashes are posted on-chain, and an administrator signs the hash batch with a private key. During execution, the contract verifies the signature against a stored public key, confirming data authenticity without exposing raw response data.

Solana's high throughput and low transaction costs make it well-suited to this use case: thousands of contributor voting events are batched into daily on-chain submissions, keeping fees negligible while preserving auditability.

The $PUBLIC token is also deployed natively on Solana as an SPL token (contract: AXCp86262ZPfpcV9bmtmtnzmJSL5sD99mCVJD4GR9vS), alongside deployments on BNB Smart Chain (BEP20) and NEAR Protocol, making it a genuinely multi-chain asset.

Token Economics

The $PUBLIC token has a maximum supply of 1 billion tokens, allocated as follows:

• 35% Community Rewards (revenue-driven issuance)
• 20% Community Distribution (airdrops and staking rewards)
• 15% Investor Funding (6-month cliff, 8% quarterly vesting)
• 15% Core Operations (10% team, 5% marketing)
• 12% Foundation Treasury
• 3% Public Sale

The most distinctive feature of PublicAI's tokenomics is its revenue-driven issuance model: community reward tokens are only minted when enterprise clients pay for datasets. The emission formula is Y = X / P, where X is the client's USD payment and P is the current $PUBLIC price. This ties new token supply directly to real economic activity rather than arbitrary schedules.

Two staking mechanisms complement this model. Collateral staking requires contributors to lock $PUBLIC as a quality bond — starting at a minimum of 100 tokens and scaling to 2x daily average earnings after day three. Collateral is returned in full if data passes consensus review and is slashed for malicious submissions. APY staking allows eligible holders to earn approximately 8% annualized returns over a six-year window, with a 14-day unstaking cooldown. Accounts flagged for data manipulation are excluded from APY staking.

Fundraising and Backers

PublicAI has raised $12 million in total funding across multiple rounds:

• A $2 million seed round closed in January 2024
• An $8 million Series A in June 2025, supported by the Stanford Blockchain Accelerator
• A $2 million CoinList public sale in August 2025, which was oversubscribed

Investors include Solana Foundation, NEAR Foundation, Blockchain Builders Fund, Foresight Ventures, Saudi Telecom Group, Everstate Capital, Taisu Ventures, and Tykhe Block Ventures. PublicAI was the first AI project featured in a CoinList sale in 2025, with its token generation event launching on Binance Alpha on August 15, 2025.

PublicAI also maintains a partnership with Chainlink, delivering high-integrity human-attested data through oracle-backed Intelligence Feeds, extending its data attestation reach into the smart contract ecosystem.

Team

PublicAI's core leadership team includes Steven Wong (CTO), Kenji Narushima (COO), Srijani Chatterjee (Head of Strategy), and Jordan Gray (Head of Marketing). The broader team includes PhDs, Stanford-affiliated researchers, and former finance professionals from Goldman Sachs and JPMorgan.

Ecosystem Fit

PublicAI enters the Solana ecosystem at a moment when AI data infrastructure is increasingly being built on-chain. The project differentiates itself through its revenue-linked token issuance, multi-tier contributor quality system, and Solana-anchored consensus recording. With over 1.2 million verified contributors, more than $14 million in reported client revenue, and live Solana smart contract deployments for RLHF consensus, PublicAI operates as an active network rather than a roadmap-stage project.