How Proximity News Network uses AI and automation
Proximity News Network Ltd may use artificial-intelligence-assisted and automated tools to support parts of its editorial, technical and administrative operations.
This Notice applies across all publications owned or operated by Proximity News Network Ltd.
Uses of AI and automation
Tools may assist with:
- monitoring public sources and news;
- identifying potentially relevant stories;
- extracting information from documents;
- transcription;
- translation;
- summarisation;
- drafting or restructuring text;
- headline and metadata suggestions;
- grammar and style review;
- categorisation, tagging and topic assignment;
- identifying duplicate material;
- matching images to stories;
- creating charts, illustrations or other visual material;
- content scheduling and distribution;
- website moderation and security; and
- analysing aggregate audience behaviour.
Not every publication or article uses every type of tool.
Editorial responsibility
AI tools are not treated as authors, editors or authoritative sources.
Proximity News Network remains responsible for the material it publishes.
The use of a tool does not reduce our obligations concerning:
- accuracy;
- fairness;
- copyright;
- privacy;
- source protection;
- defamation;
- reporting restrictions;
- sponsored-content disclosure; or
- corrections.
Verification
AI-generated output may be incomplete, biased, outdated or incorrect.
Editorial controls should therefore be proportionate to the nature and risk of the content.
In particular:
- names, dates, locations and figures should be checked;
- quotations must be traced to a reliable source;
- an AI tool must not invent a quotation;
- serious allegations require independent verification;
- legal, court, medical and financial material requires additional care;
- summaries should be checked against the underlying source; and
- generated links or citations should not be assumed to be genuine.
Human and editorial oversight
We maintain human editorial responsibility and quality controls appropriate to the content and the risks involved.
Automation may be used more extensively for routine, structured or low-risk material, but content should remain subject to systems designed to identify significant inaccuracies, legal risks and unsuitable output.
Material involving serious allegations, children, criminal proceedings, sensitive personal information or a substantial risk of harm should receive enhanced review.
Quotations and source material
Direct quotations should originate from identifiable source material such as:
- a recording;
- transcript;
- official statement;
- written response;
- court record;
- press release; or
- other verifiable source.
AI may assist with transcription or formatting, but it must not be used to create a quotation that the person did not make.
Minor corrections to obvious transcription errors may be made where they do not alter meaning.
Impersonation and deepfakes
We do not permit AI to be used to impersonate a real person deceptively or to fabricate purported evidence.
Synthetic audio, video or imagery involving identifiable people requires heightened editorial scrutiny and clear labelling where published for legitimate explanatory or journalistic reasons.
Privacy and confidential information
Staff and contractors should not enter confidential source material, unpublished sensitive information, passwords, private personal data or legally privileged material into an AI system unless:
- the use has been authorised;
- the provider and configuration have been assessed;
- appropriate contractual and security protections apply; and
- the use is necessary and proportionate.
AI providers may process information under their own technical arrangements. Approved tools should be configured to minimise unnecessary retention or reuse where controls are available.
Copyright and originality
AI-assisted content remains subject to copyright and licensing checks.
Tools must not be used knowingly to:
- reproduce a substantial part of protected material without permission or legal justification;
- imitate another publisher’s reporting and present it as original;
- remove attribution;
- circumvent paywalls or access controls; or
- generate content that unlawfully infringes another person’s rights.
Bias and discrimination
AI systems may reproduce errors or biases contained in their training data or instructions.
Editors should consider whether automated output:
- makes unsupported assumptions about an individual or group;
- uses stereotypical language;
- introduces irrelevant protected characteristics;
- disproportionately affects vulnerable people; or
- presents correlation as proof.
Disclosure to readers
We do not necessarily label every routine use of spelling, transcription, search, tagging or drafting assistance.
We may disclose AI involvement where it is material to understanding:
- how content was created;
- whether an image or recording is authentic;
- the reliability or limitations of the material;
- an experiment or automated publishing project; or
- another issue of legitimate reader interest.
Changes to our approach
AI technologies and legal standards continue to develop.
We may update this Notice as our tools and safeguards change. The current version will show its latest effective date.