Acknowledgement
Wave Particle acknowledges the Traditional Custodians of the lands on which we live and work. We pay our respects to Elders past and present, and recognise the continuing connection of Aboriginal and Torres Strait Islander peoples to land, waters, and community.
About Wave Particle
Wave Particle is a strategic communications partner for technical organisations where being misunderstood is expensive.
We work with organisations operating in complex, technical, sensitive, or regulated environments, including defence, maritime, engineering, infrastructure, and advanced industry. Our role is to close the gap between what our clients do and what their stakeholders understand.
Our work is grounded in:
- Human-led strategic thinking
- Technical empathy
- Auditable decision-making
- Disciplined continuous improvement
- Careful stewardship of confidential and sensitive information
This policy sets out how artificial intelligence may be used within Wave Particle in a way that supports those commitments rather than weakening them.
1. Introduction
Artificial intelligence is changing how communications, design, and production work are created, reviewed, and delivered. At the same time, AI introduces real risks around confidentiality, accuracy, bias, intellectual property, and over-reliance on automated systems.
Wave Particle adopts AI on the basis that production tools are becoming more accessible and more capable, but strategic meaning, judgement, and trust remain human responsibilities.
This policy explains how AI may be used within Wave Particle so that any efficiency gained at the production layer does not compromise strategic quality, client trust, or the integrity of our work.
This is not a technology-first policy. It is a strategy-first policy.
2. Purpose of This Policy
The purpose of this policy is to:
- Define how AI may and may not be used at Wave Particle
- Pprotect client confidential information, strategic reasoning, and intellectual property
- Ensure AI supports, but does not replace, human strategic judgement
- Eestablish approval, quality assurance, and accountability requirements
- Provide an ethical and operational framework for staff, contractors, and collaborators
- Support consistent adoption of AI in line with Wave Particle’s positioning, methodology, and values
3. Our Position on AI
Wave Particle recognises that AI will continue to change the baseline economics of production. As production becomes more commoditised, the differentiator shifts from output volume or polish to meaning, judgement, and stakeholder effect.
Our position is therefore simple:
- AI may support execution
- AI must not replace strategic thinking
- Client reasoning, strategic rationale, and sensitive intellectual property must remain human-controlled
- All final responsibility remains with the human author
Wave Particle does not use AI as a substitute for authorship. We use it selectively as an operational tool where it improves speed, consistency, or administrative efficiency without exposing protected information or weakening judgement.
4. Definition of AI at Wave Particle
For the purpose of this policy, artificial intelligence includes generative and assistive systems capable of producing, analysing, transforming, summarising, or classifying text, images, audio, video, code, or data.
This includes, where relevant:
- Argee language models
- Generative image tools
- Transcription and summarisation tools
- AI-assisted production and editing tools
- AI-enabled search, workflow, and automation systems
At Wave Particle, AI is defined as an assistive production and operations tool, not a strategic decision-maker.
Wave Particle’s use of AI is limited by function:
4.1 Human-Led Functions
These must remain human-led and must not be delegated to AI:
- Communications strategy
- Client positioning logic
- Stakeholder interpretation
- Value proposition development
- Workshop facilitation
- Solution design reasoning
- Framing of sensitive claims
- Interpretation of regulated, tender, or security-sensitive material
- Final judgement on meaning, tone, persuasion, or ethical suitability
4.2 AI-Assisted Functions
These may be AI-assisted where approved and appropriate:
- Formatting support
- Transcription
- Administrative summarisation
- Drafting non-sensitive internal first-pass materials
- Production assistance
- Asset variation
- Workflow automation
- Non-sensitive ideation support
- Quality support for low-risk operational tasks
5. Our AI Vision
Wave Particle’s vision is to use AI in a disciplined way that strengthens the business without diluting the trust and judgement that make the business valuable.
We believe AI can help us:
- Reduce repetitive production and administrative load
- Improve turnaround times on low-risk execution tasks
- Create more room for high-value human strategy
- Increase the frequency of refinement within our Kaizen model
- Support more consistent internal workflows
We do not believe AI should determine what our clients mean, how their stakeholders think, or what communications should ask an audience to believe.
Those decisions require context, judgement, and accountability. They remain human responsibilities.
6. Core Principles
6.1 You are the author, always
AI may assist with process or production, but the person using it remains responsible for the content, outcome, and consequences.
No work may be treated as complete merely because an AI system produced it.
6.2 Strategy stays human
All strategic reasoning, positioning logic, interpretation, and client-facing judgment must remain human-led.
AI may receive tasks or instructions. It must not receive the underlying strategic rationale where that rationale is confidential, commercially sensitive, or central to the client’s advantage.
6.3 Meaning matters more than production
Wave Particle’s value lies in translation, judgement, and stakeholder impact, not in automated output volume.
AI should be used in ways that support this distinction, not blur it.
6.4 Confidentiality comes first
Client information, particularly confidential, commercially sensitive, regulated, security-related, or tender-related information, must not be entered into unapproved AI systems.
6.5 Ethical influence over manipulation
Wave Particle applies the PCP Model, Perception, Context, Permission, as an ethical framework for communications work.
AI use must support clearer, more truthful communication. It must not be used to fabricate authority, distort meaning, or conceal material uncertainty.
6.6 Continuous improvement over convenience
AI use should strengthen Wave Particle’s Kaizen operating model by making refinement cycles faster and more informed.
Convenience alone is not enough to justify use.
6.7 Approval before adoption
No new AI tool may be used for business purposes unless it has been reviewed and approved under Wave Particle’s tool approval process
7. Approved Uses of AI
Subject to approval, confidentiality controls, and quality assurance, AI may be used for the following purposes:
7.1 Internal operational support
- Meeting note clean-up
- Task extraction from internal discussions
- Internal document formatting
- Internal workflow assistance
- Non-sensitive drafting support
7.2 Production support
- first-pass layout assistance
- asset resizing or adaptation
- transcription and subtitle generation
- non-sensitive editing assistance
- production workflow acceleration
7.3 Research support
AI may assist with early-stage research organisation where:
- The source material is non-confidential
- Outputs are independently checked
- The AI system is not relied on as an authority
- Legal, compliance, technical, or tender-critical claims are verified by a human
7.4 Internal learning and experimentation
Wave Particle may experiment with AI tools for internal research and development where:
- No client confidential material is disclosed
- Experiments are documented
- Security and IP risks are considered before use
- Outputs are not treated as approved for client use by default
8. Prohibited Uses of AI
The following uses are prohibited unless expressly approved in writing by Wave Particle leadership and supported by a specific risk review:
- Entering client confidential information into unapproved AI platforms
- Entering strategic rationale, bid logic, pricing logic, evaluation strategies, or proprietary frameworks into public or consumer AI tools
- Asking AI to determine communications strategy in place of human judgement
- Generating client-facing claims that have not been checked for accuracy
- Using AI outputs without human review
- Presenting AI-generated work as independently verified when it is not
- Using AI to create misleading authority, fabricated evidence, false testimonials, or false technical certainty
- Using AI tools that retain, train on, or repurpose business information in ways that are not approved
- Using AI tools for regulated, security-sensitive, or tender-critical work without explicit approval
- Using AI to bypass core learning, authorship, or quality expectations
9. Human-Only Strategy Rule
Wave Particle maintains a strict human-only strategy rule.
This means the following must remain outside AI systems unless expressly approved for a defined and low-risk purpose:
- Client strategy formulation
- Positioning logic
- Tender win themes
- Stakeholder persuasion architecture
- Internal reasoning behind recommendations
- Communications diagnosis where client risk is material
- Workshop synthesis involving confidential insights
- Competitive framing and response logic
- Intellectual property embedded in methods, frameworks, or strategic interpretation
AI may be instructed on the outcome of a strategic decision once that decision has been made by a human, but AI must not be used as the environment in which that strategic reasoning is exposed, developed, or stored.
10. Client Confidentiality and Intellectual Property
Client trust is central to Wave Particle’s operating model. Accordingly:
10.1 Confidential information
Client confidential information must only be used in AI systems that have been expressly approved for that purpose and assessed for:
- data retention settings
- training use restrictions
- access controls
- jurisdictional and hosting risks
- contractual protections
- auditability
10.2 Sensitive matters
Particular caution applies to:
- defence-related work
- maritime and infrastructure projects
- procurement and tender submissions
- regulated industry communications
- pricing, bid, or evaluator-facing strategy
- non-public technical material
- commercially sensitive product or service information
10.3 Intellectual property
Wave Particle’s frameworks, methods, prompts, strategic architectures, workshop materials, and client-specific reasoning are valuable intellectual property.
They must not be uploaded into any unapproved AI tool in a way that could:
- expose them to third-party model training
- weaken ownership or confidentiality claims
- make them recoverable by others
- undermine client confidence in human-led strategic work
10.4 Minimum rule
If there is doubt about whether content is safe to enter into an AI tool, do not enter it. Escalate for review first.
11. Quality Assurance and Accountability
All AI-assisted work must go through appropriate quality assurance before internal circulation or external release.
11.1 Minimum QA requirements
The person responsible for the work must review:
- consistency with the PCP Model
- factual accuracy
- context suitability
- tone and clarity
- confidentiality risk
- intellectual property risk
- alignment with strategic intent
11.2 Higher-risk work
Higher-risk work requires additional review. This includes:
- tender responses
- defence or security-adjacent communications
- investor-facing claims
- regulated industry content
- legal, compliance, or technical claims
- sensitive stakeholder communications
11.3 Accountability
The human reviewer remains accountable for the output, even where AI materially assisted its production.
AI assistance does not reduce the reviewer’s responsibility.
14. Tool Approval and Security
Only approved AI tools may be used for Wave Particle work.
Before any tool is approved, Wave Particle should consider:
- what data the tool stores
- whether prompts or files are retained
- whether inputs may be used for model training
- who can access the data
- whether enterprise controls are available
- whether deletion is possible
- where data may be hosted
- whether the tool creates IP or licensing uncertainty
- whether the supplier terms are suitable for business use
Approval should be documented and reviewed periodically.
If a tool changes its terms, security model, or training settings, approval should be reassessed.
15. Training and Capability
Everyone using AI within Wave Particle must be trained to use it appropriately.
Training should cover, at a minimum:
- this policy
- confidentiality and security obligations
- prompt hygiene
- hallucination and inaccuracy risk
- bias and fairness risk
- intellectual property risk
- when to escalate or avoid AI entirely
- how AI fits within Wave Particle’s methodology and positioning
No person should use AI in client work unless they understand both the tool’s limitations and the business rules that apply to it.
16. Risk Areas and Mitigation Measures
[E.g., "Wave Particle will design and produce a 60-second brand video for ABC Corp's product launch campaign, including concept development, scripting, shot list, production, and post-production (colour grade, sound design, titles)."]
16.1 Confidentiality risk
Risk: Client information is disclosed to an inappropriate tool or environment.
Mitigation: Approved tools only, minimal data entry, human-only strategy rule, escalation where uncertain.
16.2 Accuracy risk
Risk: AI produces false, incomplete, or misleading outputs.
Mitigation: Mandatory human verification, especially for technical, legal, tender, or client-facing material.
16.3 Intellectual property risk
Risk: Client IP or Wave Particle IP is exposed, diluted, or incorporated into third-party systems.
Mitigation: No upload of strategic reasoning or protected frameworks into unapproved tools; review tool terms before use.
16.4 Ethical risk
Risk: Communications are designed in a way that overstates, manipulates, or creates false authority.
Mitigation: PCP Model review, truth-grounded framing, human judgement over persuasion decisions.
16.5 Overdependence risk
Risk: Team capability erodes because AI is used as a substitute for thinking.
Mitigation: Human authorship rule, QA discipline, training, and retaining core craft skills.
16.6 Reputational risk
Risk: Poor or inappropriate AI use damages client trust or Wave Particle’s market position.
Mitigation: Conservative use in high-trust environments, transparency where appropriate, and documented internal controls.
16.7 Security risk
Risk: AI tools introduce vulnerabilities or unauthorised access pathways.
Mitigation: Tool approval, access control, password hygiene, periodic review, and removal of unapproved tools.
16.8 Compliance risk
Risk: AI use creates issues under privacy, consumer, copyright, contractual, or industry-specific obligations.
Mitigation: Human review, contract-aware use, and external legal input where the matter is sensitive or unclear.
17. Governance and Review
This policy applies to all Wave Particle personnel, including employees, contractors, consultants, and collaborators who use AI in connection with Wave Particle work.
17.1 Ownership
The policy owner is Wave Particle leadership.
17.2 Review cycle
This policy should be reviewed:
- at least every 12 months
- when a major new AI tool is adopted
- when client requirements materially change
- after any material incident involving AI use
- when legal or regulatory developments require updates
17.3 Exceptions
Any exception to this policy must be:
- approved in advance
- limited in scope
- documented
- supported by a clear risk rationale
18. Final Statement
Wave Particle uses AI from a position of discipline, not novelty.
We accept that production tools will continue to improve and become more common. We do not treat that as a threat to our identity. We treat it as a reason to be even clearer about where our value sits.
Our value is in human judgement, technical empathy, strategic translation, ethical communications design, and a system that improves with every cycle.
AI may help us execute. It does not define what we think, what we mean, or what we stand behind.
That remains ours.
