Responsible AI Usage Policy

Aperim Pty Ltd (ABN 46150699737; ACN 150699737) is incorporated in New South Wales, Australia and operates the aicial brand. In this policy, “aicial” refers to that brand; “we”, “us” and “our” refer to Aperim Pty Ltd; and “you” refers to the person or organisation reading this policy, including a customer, prospective customer or other individual affected by our use of artificial intelligence.

We describe our commitments about artificial intelligence in more than one place, including in product and marketing material. This Responsible AI Usage Policy is the operative legal statement of those commitments; where another description of our AI practices is inconsistent with it, this policy governs.

This policy applies to our website, waitlist and services. Those services include fixed-scope social-media audits, monthly reporting retainers and benchmark studies, as well as the aicial technology used to deliver them, and it applies to our planned self-serve software service once it is made available. From the date that service is available, the artificial intelligence within it is governed by this policy on the same basis as our current services.

Artificial intelligence is used across these services, including to generate cross-platform performance analytics and reports; to draft and schedule social content based on a customer’s own top-performing material and instructions; to detect likely impersonation and deepfake risk affecting a customer’s identity; to produce social-listening insights and benchmark analysis; and to power customer-facing automation such as care and support replies. The AI Usage & Disclosure Matrix lists every AI-enabled feature individually, together with what each system does and does not do.

These principles govern every AI system aicial builds, deploys, or relies on a provider to supply. They apply regardless of the model, provider or feature involved, and they take precedence over any product requirement that would conflict with them.

• Consent before collection. Data a customer contributes beyond what their own paid service needs — for example, to our benchmark panel — is collected only under clear, unbundled, plain-language opt-in, never bundled into account sign-up. You can withdraw that consent at any time, and we propagate the withdrawal, and the resulting deletion, end-to-end, including to our service providers. Processing a customer’s own account data to deliver the service itself — for example, to generate their analytics and reports — relies instead on the contract and legitimate-interest bases our Privacy Policy describes.
• Aggregate before individual. Insight drawn from data beyond a customer’s own connected accounts is presented in aggregate, with minimum group-size and other disclosure-limitation safeguards, not at the level of an identifiable individual. Profile-level analysis is limited to a customer’s own connected accounts, and to public figures where we have obtained legal advice supporting that analysis.
• No manipulation of platforms or people. We do not build or offer fake engagement or follower acquisition in any form; engagement pods or reciprocal-engagement schemes; coordinated inauthentic behaviour; sockpuppet or astroturfed accounts or content; automation that poses as a human without disclosure; or mass-messaging, auto-follow or comment-spam tooling sold as “growth”. We do not build features designed to evade a platform’s anti-abuse systems. Every optimisation feature we ship is confined to a customer’s own honest content, timing and format choices — a real, revenue-generating category of features is closed to us by this commitment, and we accept that on purpose.
• Label what AI made. Every synthetic output aicial generates carries machine-readable provenance marking, and AI-driven automation discloses itself. Section 5 states this commitment in full.
• Honest numbers only. Every benchmark we publish states its sample size. Where our data has coverage limits, we say what we cannot see. We do not present a real figure with more exactness than our data supports. A modelled or estimated figure is labelled as modelled or estimated, in the product and in our marketing, and protection-related marketing uses only real, customer-consented incidents.
• No circumvention. We do not bypass a platform’s technical access controls — CAPTCHAs, JavaScript challenges, bot defences or rate limits — anywhere in our data-collection chain, and every provider who collects data on our behalf warrants the same to us contractually. We honour robots directives and opt-out registries.
• No biometric search, no data brokering. We do not run face- or voice-embedding searches over scraped or third-party media. We do not resell an individual’s data at the row level. Commerce and deal data we hold is used only to serve the contributing customer and to produce the aggregate benchmarks described in this policy and our Privacy Policy.
• You own your data. A customer can export their data at any time and leave cleanly. We do not build hostage mechanics into any aicial service.
• Sell outcomes honestly. We do not promise to fix a platform’s payout algorithm, restore organic reach, or “beat” a platform’s algorithm. We market diversification, diagnosis and evidence, not miracles, including in marketing directed at a customer experiencing income instability.
• Meet the higher bar. Where law and our own ethics point in different directions, or where we are genuinely unsure, we choose the more protective course: aggregate the data, disclose the practice, or do not ship the feature.

We voluntarily align these principles with the Australian Government’s Guidance for AI Adoption, published by the National AI Centre in late October 2025 to replace the earlier Voluntary AI Safety Standard. It sets out six practices for responsible AI governance — deciding who is accountable, understanding impacts, measuring and managing risks, sharing essential information, testing and monitoring, and maintaining human control — and it is voluntary, non-binding guidance rather than legislation. We treat it as the current Australian good-practice benchmark for how we govern AI, without treating voluntary alignment as a certification or as compliance with any future mandatory scheme.

Accountability for this policy, and for how artificial intelligence is used across aicial’s services, sits with Aperim Pty Ltd’s leadership today. As the company and the risk profile of the AI systems we operate grow, we will formalise that oversight — clearer internal roles, a defined review process and regular reporting — proportionate to that growth, and we will update this policy when we do.

This policy binds every aicial employee, contractor and provider who builds or operates any part of an aicial AI system. Anyone unsure whether a feature or a provider complies with this policy is expected to raise it before shipping the feature or engaging the provider, not after.

We are voluntarily aligning our practices with the Influencer Marketing Code of Practice published by the Australian Influencer Marketing Council (AiMCO), the Australian industry body for influencer marketing, and we intend to pursue AiMCO accreditation. We have not yet obtained that accreditation, and we will update this policy when we do.

We classify every AI-enabled feature we build against the four-tier risk framework used to describe the European Union’s AI Act (Regulation (EU) 2024/1689) — unacceptable, high, limited and minimal risk. We use this framework because it is external, well understood and independently legible to a customer, regulator or investor assessing our practices — not because every aicial feature is subject to EU jurisdiction.

We do not build or deploy a system that would fall within the unacceptable-risk tier, regardless of jurisdiction or how a regulator would classify them.

Where a feature’s classification changes as the feature, its use or the applicable law changes, we reclassify it and adjust our controls and disclosures accordingly. The current classification of each AI-enabled feature we operate is set out in the AI Usage & Disclosure Matrix.

We apply the disclosure standard set by Article 50 of the EU AI Act (Regulation (EU) 2024/1689), which applies from 2 August 2026, to every user-facing AI interaction across our services and in every market we operate in — not only where a law requires it. Where you are interacting with an AI system and that would not otherwise be obvious, we tell you. Once our planned self-serve software service is available, its in-product assistant will disclose that it is an AI system at your first interaction with it in a session, regardless of whether that would otherwise be obvious.

We apply machine-readable content-provenance marking — of the kind defined by the Coalition for Content Provenance and Authenticity (C2PA) — to every synthetic output aicial generates, built to withstand ordinary platform re-encoding, so the origin of that content can be checked downstream.

Where a customer-facing interaction is automated — including automated care and support replies — we disclose that it is automated.

Our Privacy Policy explains how we collect, minimise, secure, retain and, where applicable, delete the personal information our AI systems use and produce, including its section on AI and personal information specifically. This policy does not repeat those commitments. Read together, neither policy narrows the protection the other provides.

The principles in section 2 — in particular consent before collection, aggregate before individual, and no biometric search or data brokering — apply to every AI system regardless of which policy states the operational detail behind them. Where a provider trains or fine-tunes a model using aicial customer data, the limits our Privacy Policy places on cross-customer model training apply to that training.

We do not let an AI system make a solely automated decision that produces a legal or similarly significant effect on a person. A person genuinely reviews any consequential AI output before we act on it — including every adverse identity-protection finding before we take an external step such as a takedown request — and that review is not a formality.

If you believe an AI-influenced output about you or your account is wrong, or you want a person to review something an aicial AI system produced, contact ethics@aicial.com. Where applicable law gives you a formal right to contest an automated decision, obtain human intervention or express your view, our Privacy Policy explains how to exercise it.

Where a provider collects data for us, supplies a model, or performs any part of an AI system on our behalf, the principles in section 2 apply to that provider as they apply to us. We contractually require a provider who collects data on our behalf to warrant that it does not circumvent a platform’s technical access controls, consistent with the No circumvention principle in section 2, and we may audit or terminate a provider that breaches that warranty.

We assess an AI provider’s own practices — including its data handling, security and, where relevant, its own disclosure practices — before we rely on it, and periodically afterwards. The Sub-processor List identifies the providers that may process customer data, including those that provide AI functionality.

This policy states the principles, governance and disclosure standard that apply to every AI system aicial operates. The AI Usage & Disclosure Matrix is its technical companion: for each AI-enabled feature, it names the system, states what it does and does not do, its classification under section 4, the data it uses, and the disclosure a user sees.

Where this policy and the AI Usage & Disclosure Matrix appear to differ, this policy governs.

We review this policy, our risk classifications and the AI Usage & Disclosure Matrix at least once a year, and sooner if we launch a materially new AI-enabled feature or a law that governs us changes. The version, effective date and last-reviewed date shown at the top identify the current policy.

We publish an updated policy on our website and, where a change is material, provide additional notice through the service or by email before it takes effect.

Australia’s Guidance for AI Adoption is non-binding. If it is replaced or strengthened, or if a law that governs us is amended to address artificial intelligence specifically, we will update this policy and our compliance posture to reflect it.

For questions about this policy, to raise a concern about how we use artificial intelligence, or to ask a person to review an AI-influenced output, contact:

Aperim Pty Ltd, which operates the aicial brand
ABN 46150699737
ACN 150699737
New South Wales, Australia
Email: ethics@aicial.com