AI Use Disclosure

last updated:

Arkhouse is an AI optimised agency. We use artificial intelligence to work faster, research deeper and test more ideas, and we do it under human direction, judgement and accountability. This disclosure explains how we use AI, the rules we hold ourselves to, and what it means for clients and for people who see the work we make. It forms part of our Terms of Service and Privacy Policy.

1. Our principles

  • Humans decide. AI assists our strategists, designers, writers and analysts. Every strategy, piece of creative and recommendation is directed, reviewed and approved by a person at Arkhouse before it reaches a client or the public.

  • We tell you. Clients are told that AI is used in our process, and are told on delivery when a specific asset is AI generated or materially AI manipulated.

  • We label what could mislead. Synthetic imagery, voice or video that could be mistaken for a genuine depiction of real people, places or events is labelled as artificially generated when it is published.

  • We protect your data. Confidential client information and personal information are not entered into AI tools that train on inputs or that lack contractual confidentiality and data protection terms.

  • We do not automate judgements about people. We do not use AI to make decisions about individuals that have legal or similarly significant effects.

  • We respect creators. We do not prompt AI tools to imitate the identifiable style of a named living artist, designer or writer, or to reproduce protected characters, logos or works.

2. How we use AI


Area

Typical uses

Human role

Research and strategy

Market, competitor and audience research synthesis, trend scanning, data analysis, summarising transcripts

Strategist verifies sources, checks figures and owns conclusions

Copy and content

First drafts, variations, tone adaptation, translation support, SEO analysis

Writer edits, fact checks and approves; brand voice is set by people

Design and production

Concept exploration, image generation and extension, background removal, retouching, resizing, layout variations, motion assistance

Designer directs, selects, refines and signs off

Advertising and media

Audience and bid optimisation inside ad platforms, performance analysis, creative testing

Media lead sets strategy, budgets and guardrails

Operations

Meeting transcription and summaries, project documentation, reporting

Team reviews and corrects before sharing

3. What this means for clients

AI use is covered by clause 14 of our Terms of Service. In summary: you may ask us to exclude AI from specific tasks; you may ask for a record of the AI tools used on your work; AI generated or materially manipulated assets are identified to you on delivery; and intellectual property in final deliverables is assigned to you on payment in the normal way. Because copyright protection for purely AI generated output is uncertain in many jurisdictions, including South Africa, the European Union and the United States, we tell you when a hero asset is substantially AI generated so that you can decide whether human authored alternatives are needed for registration or enforcement purposes.

4. Meeting recordings

We use AI meeting assistants to record and transcribe some client meetings so that briefs and decisions are captured accurately. We tell participants at the start of a recorded meeting and stop recording on request. Recordings and transcripts are stored under the retention periods in our Privacy Policy and are not used to train AI models.

5. Personal information and AI tools

We select AI providers under written terms that prohibit training on our inputs, provide confidentiality, and offer data processing terms that meet POPIA, the GDPR and the UK GDPR. Where an AI tool processes personal information on behalf of a client it is listed as a subprocessor in our Data Processing Addendum. We do not upload client customer databases, research respondent data or sensitive material to general purpose consumer AI tools.

6. Labelling and the law

South African law does not yet contain a general obligation to label AI generated content, but the Advertising Regulatory Board’s Code requires advertising to be honest and not misleading, and POPIA section 71 restricts decisions based solely on automated processing. Where work is intended for, or foreseeably reaches, audiences in the European Union, Article 50 of the EU Artificial Intelligence Act applies from 2 August 2026: deployers must disclose that image, audio or video content constituting a deepfake has been artificially generated or manipulated, and must disclose AI generated text published to inform the public on matters of public interest unless it has undergone human editorial review with a person holding editorial responsibility. Content that is evidently artistic, creative, satirical or fictional need only disclose the existence of the generated content in a way that does not hamper the work. We meet these standards for all clients as a matter of policy, regardless of the audience’s location.

7. What we do not do

We do not create deepfakes of real people without their documented consent and clear labelling. We do not generate fake reviews, testimonials, endorsements or user generated content presented as genuine. We do not use AI to scrape personal information or to build marketing lists without a lawful basis. We do not use emotion recognition or biometric categorisation systems.

8. Questions

If you have questions about how AI was used in a piece of work, or you believe content we published is mislabelled, email connect@arkhouse.co.za with the subject line AI disclosure. We review this disclosure at least annually and as our tools and the law evolve.

Your Whole Marketing Team, One Call Away.

Tell us what you need in 30 minutes. You’ll get a clear plan and a fixed price in writing, with no obligation.