Sustainable data products and practices

Instruction and application

Complete

Sustainable AI is not only about watts per FLOP. Data products—features, tables, labels and pipelines—shape storage, recomputation and risk. Lean, well-governed data reduces cost, emissions and harm.

Data minimisation

Collect and retain only what you need. Over-collection increases storage energy, widens breach impact and complicates consent. Minimisation aligns with GDPR-style principles and often improves model focus when paired with strong definitions.

Practices include selective retention, aggregation for analytics where raw detail is unnecessary, and anonymisation or pseudonymisation when identifiers are not required.

Data quality and integrity

Sustainable models need trustworthy inputs: validation rules, automated checks, versioned datasets and clear ownership. Poor quality forces retraining, hotfixes and rework—all of which burn energy and trust.

Responsible data governance

Governance connects policy to implementation:

✨

Access control

limits sensitive data to authorised roles.

✨

Audit trails

show who changed what and why—critical for regulated domains.

✨

Bias and fairness assessments

catch discriminatory patterns before they scale; see Algorithmic Justice League and IBM discussion of real-world bias.

Energy-aware algorithms and hardware

Pair lean data with efficient computation: smaller models where adequate, quantisation, specialised chips and vendor tools that deduplicate storage (for example NetApp-style dedup/compression patterns in enterprise storage).

Key point

Sustainability wins stack: less data movement + higher data quality + more efficient compute beats optimising only one layer.

Action item: Quiz

A company is developing a new AI-powered personalised advertising platform. It has access to a wide range of user data, including browsing history, purchase records, social media activity and location data. To minimise the environmental impact and enhance user privacy, what data strategy should the company prioritise?

A) Collect and store all available user data to maximise the platform's personalisation capabilities.B) Retain all user data indefinitely to enable future improvements to the platform's algorithms.C) Anonymise and aggregate user data whenever possible, and only collect and retain the data that is strictly necessary for ad personalisation (e.g. purchase history and browsing behaviour related to specific product categories).D) Share the raw, unanonymised user data with third-party advertising partners to increase revenue potential.

A financial institution is using machine learning to automate loan application processing. They discover that their historical loan data contains inaccuracies and inconsistencies, such as missing income information for certain applicants and inconsistent formatting of employment history. To ensure the AI system is both sustainable and trustworthy, what should they do?

A) Proceed with training the model on the existing data, as cleaning the data would be too time-consuming and costly.B) Remove the applicants with incomplete or inconsistent data from the dataset to ensure data quality.C) Implement data validation techniques to detect and correct inaccuracies and inconsistencies in the data, and establish automated data cleaning processes for ongoing data integrity.D) Manually correct only a small subset of the data and assume that the model will generalise to the remaining errors.

A healthcare company is developing an AI system to analyse patient records and predict the risk of developing certain diseases. To ensure responsible data governance and maintain patient trust, what practices should the company implement?

A) Allow all researchers and developers within the company to access the raw patient data to facilitate innovation.B) Use the patient data for purposes beyond medical diagnosis, such as marketing pharmaceuticals, without obtaining explicit patient consent.C) Implement stringent data access policies, maintain detailed audit trails of data usage, and conduct bias detection and fairness assessments to ensure the AI system operates transparently, ethically and in compliance with GDPR regulations.D) Store all patient data in a centralised database without any access controls or audit trails to simplify data management.

In progress