ਕਾਨੂੰਨੀ
AI ਖੁਲਾਸਾ
ਪੂਰੀ ਨੀਤੀ ਅੰਗਰੇਜ਼ੀ ਵਿੱਚ ਹੈ ਅਤੇ ਕਾਨੂੰਨੀ ਤੌਰ ’ਤੇ ਉਹੀ ਲਾਗੂ ਹੁੰਦੀ ਹੈ। ਹੇਠਾਂ ਦਿੱਤਾ ਪੰਜਾਬੀ ਸੰਖੇਪ ਸਹੂਲਤ ਲਈ ਹੈ; ਇਹ ਨੀਤੀ ਦੀ ਥਾਂ ਨਹੀਂ ਲੈਂਦਾ।
Effective — to supply —
ਸੰਖੇਪ
- ਹਰ ਲਿਖਤ ਦੀ ਜ਼ਿੰਮੇਵਾਰੀ ਕਿਸੇ ਇਨਸਾਨ ਸੰਪਾਦਕ ਦੀ ਹੈ। ਕੋਈ ਸੰਦ ਇਹ ਨਹੀਂ ਤੈਅ ਕਰਦਾ ਕਿ ਅਸੀਂ ਕੀ ਛਾਪੀਏ, ਅਤੇ ਕੁਝ ਵੀ ਬਿਨਾਂ ਵੇਖੇ ਨਹੀਂ ਜਾਂਦਾ।
- AI ਸੰਦ ਅਸੀਂ ਸਿਰਫ਼ ਮਦਦ ਲਈ ਵਰਤਦੇ ਹਾਂ: ਅਨੁਵਾਦ ਦੇ ਖਰੜੇ, ਆਡੀਓ ਤੋਂ ਲਿਖਤ, ਤਸਵੀਰਾਂ ਦੀ ਤਕਨੀਕੀ ਸੋਧ ਅਤੇ ਖੋਜ, ਹਮੇਸ਼ਾ ਇਨਸਾਨੀ ਨਿਗਰਾਨੀ ਹੇਠ।
- ਅਸੀਂ AI ਨਾਲ ਲਿਖੀਆਂ ਲਿਖਤਾਂ ਇਨਸਾਨ ਦੇ ਕੰਮ ਵਜੋਂ ਨਹੀਂ ਛਾਪਦੇ, ਝੂਠੇ ਲੇਖਕ ਨਹੀਂ ਘੜਦੇ, ਅਤੇ AI ਨਾਲ ਬਣੀ ਤਸਵੀਰ ਨੂੰ ਅਸਲ ਫ਼ੋਟੋ ਵਜੋਂ ਪੇਸ਼ ਨਹੀਂ ਕਰਦੇ।
- ਜਿੱਥੇ AI ਦੀ ਮਦਦ ਨੇ ਕਿਸੇ ਲਿਖਤ ਨੂੰ ਸੱਚਮੁੱਚ ਘੜਿਆ ਹੋਵੇ, ਉੱਥੇ ਅਸੀਂ ਸਾਫ਼ ਲੇਬਲ ਲਾਉਂਦੇ ਹਾਂ, ਅਤੇ ਉਹ ਲੇਬਲ ਇਸੇ ਸਫ਼ੇ ਨਾਲ ਜੁੜਦਾ ਹੈ।
1. Principles
Northview Productions Inc., a British Columbia company doing business as Kalam Di Goonj (“the Company”, “we”), publishes Punjabi-first writing. Two principles govern everything below:
- A human is always responsible. Every published piece has an accountable human editor behind it. No tool decides what we publish, and nothing is published unreviewed. Errors are ours to correct under our Corrections Policy, whatever tool was involved.
- The craft is human. The poetry, the reporting, and the Punjabi prose readers come for are written by people. Tools may support that work; they do not replace it.
2. Where AI may assist
We may use AI-assisted tools, under human control, for:
- Translation drafts. A machine draft of an English or other rendition may serve as a starting point, always reviewed, corrected, and completed by a human translator before publication. The canonical Gurmukhi work is human.
- Transcription. Converting audio from interviews, narration, and oral histories into text. Transcripts of record are checked by a human before we rely on or publish them, and oral-history transcription receives particular care.
- Image processing. Technical work on photographs and artwork, including restoration of damaged archival images, colour and exposure correction, resizing, and alt-text drafting. Not the generation of scenes.
- Research assistance. Background searching, summarizing public documents, and locating sources. Anything that reaches publication is verified by a human against original sources. A model’s output is never a source.
- Operational filtering, including spam filtering and comparable infrastructure that keeps inboxes and future comments usable.
3. Where AI is not used
- No AI-generated articles presented as human work. We do not publish machine-written pieces under a human byline, and we do not present generated text as reporting or literature.
- No fabricated authors. Every byline belongs to a real person. Pen names of real people are a separate and disclosed tradition. There are no synthetic contributors.
- No AI-generated imagery presented as documentary photography. A generated or materially AI-altered image is never presented as a photograph of a real event, person, or place. Where a generated illustration is used as illustration, it is labelled as such.
4. Labelling
Where AI assistance was material to a published piece, for example a rendition built from a machine draft, a substantially machine-transcribed oral history, or a labelled generated illustration, the piece carries a note saying so, and the note links to this page. Routine, non-material uses such as spell-check-grade assistance, spam filtering, and search are covered by this page and are not labelled piece by piece. The Editor-in-Chief decides borderline cases, and our standing instruction is to label rather than omit.
5. Contributors’ duty of disclosure
Contributors must disclose material AI assistance in submitted or commissioned work at the time of submission, under our Contributor Submission Terms. Submitting AI-generated work as one’s own writing is a serious breach and may result in rejection, withdrawal of published work, and termination of the relationship.
6. Your data and AI tools
We do not put readers’ personal information into third-party AI tools without disclosing it. Where a tool that processes personal information, for example transcription of an interview you gave, involves a third-party AI service, that processing is disclosed in our Privacy Policy like any other processor, with the same PIPEDA and PIPA treatment, including where the data goes and how long it is kept. Reader accounts, subscription records, and correspondence are not used to train any model.
7. Limits of this page
- It does not claim the Publication is AI-free. That claim would be neither accurate nor stable, since ordinary software increasingly incorporates these tools.
- It does not qualify our accuracy commitments. A tool’s error published by us is our error, correctable under our Corrections Policy.
- It does not restrict what we report about AI, which is journalism like any other subject.
8. Review
Tools, norms, and law are all changing. We review this page every six months and whenever we adopt a tool that touches content or reader data, and we record changes of practice here.