Sustainability Statement
(Environmental Responsibility & AI Use)
1. Purpose
This statement sets out askKira’s approach to environmental sustainability in the design, deployment, and ongoing operation of its AI systems.
It recognises that AI systems, particularly those based on large language models, have material environmental impacts through compute usage, data storage, and infrastructure dependencies. askKira is committed to proportionate, transparent, and improving practice over time.
2. Scope
This statement applies to:
The askKira AI platform
Associated cloud infrastructure and hosting
Product design decisions that influence compute demand
It does not currently extend to:
Full lifecycle assessment of third-party foundation models
Upstream manufacturing impacts of hardware outside askKira’s control
3. Design Principles
askKira’s sustainability approach is guided by the following principles:
Proportionality – matching compute use to educational value
Efficiency – avoiding unnecessary processing or duplication
Transparency – being clear about what is known and unknown
Continuous improvement – measuring and improving over time
4. Infrastructure & Hosting
Hosting approach:
askKira uses reputable cloud infrastructure providers operating within the UK or equivalent regulatory environments.
Environmental considerations:
Reliance on providers that publicly commit to carbon reduction, energy efficiency, and increased use of renewable energy.
Infrastructure is shared and virtualised, avoiding dedicated on-premise hardware where possible.
Gaps / to confirm:
☐ Named hosting and cloud providers
☐ Data centre region(s)
☐ Provider-specific sustainability credentials
5. Product & System Design Choices
askKira seeks to reduce unnecessary environmental impact through:
Task-appropriate model use
Avoiding over-specification where simpler responses suffice.
Rate limiting and usage controls
Preventing excessive or abusive compute consumption.
Session efficiency
Encouraging concise, purposeful interactions rather than repeated trial-and-error prompting.
No model training on customer data
Avoids additional compute-intensive retraining cycles.
Gaps / to confirm:
☐ Explicit small-model vs large-model routing strategy
☐ Caching or reuse mechanisms
☐ Monitoring of average compute per interaction
6. Measurement & Monitoring
askKira does not currently claim full carbon accounting for AI usage. Instead, it commits to progressive measurement.
Initial metrics under consideration include:
Volume of AI interactions over time
Relative compute intensity per interaction (proxy metrics)
Infrastructure efficiency indicators provided by cloud suppliers
Over time, askKira aims to:
Improve visibility of compute and energy use
Track trends rather than absolute precision
Use data to inform product design decisions
Gaps / to confirm:
☐ Baseline year for measurement
☐ Reporting cadence
☐ Customer-facing sustainability metrics
7. Supplier & Supply Chain Considerations
askKira recognises that environmental impact is significantly influenced by third-party suppliers.
Approach:
Preference for suppliers with published sustainability commitments.
Periodic review of supplier posture where commercially and technically feasible.
askKira does not currently conduct independent audits of supplier environmental performance.
Gaps / to confirm:
☐ Supplier review criteria
☐ Thresholds for supplier acceptability
☐ Disclosure of sub-processors in sustainability context
8. Governance & Accountability
Environmental sustainability is treated as part of askKira’s broader ethical and responsible AI governance.
Governance approach:
Sustainability considerations included in major product or infrastructure decisions.
Responsibility sits with senior leadership rather than being fully delegated to technical teams.
Gaps / to confirm:
☐ Named sustainability or ethics owner
☐ Formal review cycle
☐ Board-level oversight arrangements
9. Relationship to the UK Government Data & AI Ethics Framework
This statement supports the Environmental Sustainability principle within the UK Government Data & AI Ethics Framework by:
Acknowledging environmental impacts
Taking proportionate steps to reduce unnecessary compute
Being transparent about current limits and future intent
10. Forward Commitments
askKira commits to:
Avoid overstating environmental performance
Improve transparency as measurement capability matures
Balance innovation in education with responsible resource use
This statement will be reviewed and updated periodically as the platform, data availability, and industry standards evolve.
Summary for Buyers
askKira takes a measured and credible approach to AI sustainability:
No greenwashing
Clear acknowledgement of impact
Practical steps to reduce unnecessary compute
Commitment to improvement over time
The gaps identified represent areas for future development, not unmanaged risk, and reflect current best practice for proportional AI governance.
