Dr. Liam Whitaker

University of Sydney Study: AI's Impact on Australian News Media – Copilot Sidelining Local Journalism

Key Findings: AI Amplifies Global Voices Over Australian Sources

university-of-sydneyai-studyaustralian-journalismmicrosoft-copilotnews-media-bargaining-code

See more Higher Ed News Articles Join the Conversation!

Key Findings from the University of Sydney Study on AI and Australian News Media

The University of Sydney's latest research has uncovered a troubling trend in how artificial intelligence (AI) tools are influencing the way Australians access news. Led by Dr. Timothy Koskie, a Post-Doctoral Research Associate in Digital Communication at the university's Centre for AI, Trust and Governance, the study reveals that Microsoft Copilot – a popular AI chatbot integrated into Windows and Bing – systematically sidelines Australian journalism in its news summaries. 0 71 In an analysis of 434 AI-generated responses, local sources appeared in only about one-fifth of the outputs, with global giants like CNN and BBC dominating the landscape. This bias not only amplifies international voices but also erases the contributions of Australian journalists and regional stories, posing risks to media diversity and democratic discourse.

Dr. Koskie's work, detailed in the Australian Journalism Review and highlighted in recent university announcements, emphasizes that AI platforms are "replicating and exacerbating existing power imbalances" in Australia's media ecosystem. 69 With concentrated media ownership and growing news deserts in regional areas, this development could accelerate the decline of independent outlets, reducing traffic and revenue as users consume summarized content without visiting original sites.

Unpacking the Methodology: Testing Copilot in an Australian Context

To expose these patterns, Dr. Koskie conducted a rigorous 31-day experiment using Microsoft Copilot configured for an Australian user. He inputted seven globally focused prompts recommended by the platform itself, such as "What are the major health or medical news updates for this week?" and "What are the top global news stories today?" Each prompt generated multiple summaries, totaling 434 outputs for analysis. 70

The focus was on source representation, attribution of voices, and geographical biases rather than factual accuracy or misinformation. Results showed over half of the most-referenced websites were US-based, with European outlets like the BBC also prominent. Australian media links appeared in roughly 20% of responses, and shockingly, none in three out of the seven prompts. Even when included, they skewed toward major players like ABC and Nine, marginalizing independents and regionals. 68

  • 434 total AI-generated summaries examined.
  • 7 standardized prompts used consistently.
  • Australian sources in ~20% of outputs.
  • 0 Australian sources in 43% of prompt categories (3/7).

This methodology highlights how AI, despite location tagging, defaults to internet-dominant sources, homogenizing journalists as generic "researchers" or "experts" and stripping local context – rarely mentioning places like Ballarat or the Kimberley.

Graph showing source bias in Microsoft Copilot news summaries from University of Sydney study

Real-World Examples: What Copilot Delivers to Australian Users

Consider a prompt on global health news: Copilot's response hyperlinks primarily to US sites like USA Today, with Australian outlets absent. Journalists' names vanish, replaced by vague attributions, and Australian-specific angles – vital for public health policy – are overlooked. In political news summaries, CNN and BBC dominate, introducing non-local perspectives like ABC America despite the user's Aussie location. 70

Dr. Koskie notes, "Australians are invisible in this. In international studies, what people trust is the local news." This erasure extends to communities, fostering a "one-way relationship" where platforms extract value without compensating local media. For academics and media researchers exploring AI ethics, these cases underscore the need for transparent algorithms that prioritize cultural relevance.Explore research positions in media studies at Australian universities.

Australian Media Landscape: Pre-Existing Vulnerabilities Amplified by AI

Australia's news sector already grapples with challenges: high ownership concentration (e.g., News Corp and Nine controlling vast shares), declining ad revenue, and news deserts affecting over 100 regional titles shuttered in recent years. AI summaries exacerbate this by diverting traffic – users get instant digests without clicks, starving outlets of essential revenue. 2

The University of Sydney study warns of deepened inequalities, with independents and regionals hit hardest. Reuters Institute's Digital News Report 2025 echoes concerns, noting AI chatbots as emerging news sources amid stagnating trust in traditional media. 57

Positive Side: AI Adoption Boosting Efficiency in Australian Newsrooms

Not all AI impacts are negative. News Corp Australia has pioneered tools like NewsGPT since 2023, automating local story generation and transcription to free journalists for investigative work. In 2025, they expanded generative AI for routine tasks, treating it as a "productivity aid" while upholding editorial oversight. 20

  • Speeding transcription and drafting.
  • Analyzing public records faster.
  • Enabling more enterprise reporting.

UTS research shows cautious uptake, with newsrooms balancing efficiency gains against ethical risks. For higher ed professionals in journalism programs, this dual role of AI – disruptor and enabler – offers rich study material. Career advice for research assistants in media AI.

Democratic Risks: News Deserts, Trust Erosion, and Invisible Voices

Local news is "democratic infrastructure," per Dr. Koskie. AI's sidelining risks more news deserts, fewer diverse voices, and eroded trust – Australians favor local sources, yet AI pushes global ones. Without journalists named or regions highlighted, public discourse homogenizes, weakening civic engagement on issues like climate or health policy tailored to Aussie contexts.

Broader ripple: reduced funding leads to layoffs, further consolidation. The study predicts a "weakened democracy" without checks.Guardian coverage details these threats.

Policy Pathways: Extending the News Media Bargaining Code to AI

Dr. Koskie advocates extending Australia's News Media Bargaining Code – which forced Google and Meta to pay $250M+ annually to publishers – to AI tools. Mandate geo-location coding to surface local sources and assess platforms' impact via bargaining incentives.

  • Recognize local news as essential infrastructure.
  • Collaborate with AI firms on public-interest design.
  • Regulate summaries as news services.

The ACCC could designate Copilot-like tools, fostering fair deals. Internationally, similar calls grow amid Reuters 2026 predictions of AI upending news.

Expert Perspectives and Stakeholder Views

"AI is preferencing dominant international sources and sidelining independent and regional media," says Koskie. Media execs worry about revenue; ethicists highlight pluralism erosion. Balanced views note AI's potential if regulated – e.g., UTS studies on gen AI attitudes show optimism with safeguards.

For university researchers, this intersects AI governance and comms studies. Lecturer roles in digital media at Australian unis.

Future Outlook: Balancing Innovation with Local Priorities

By 2026, AI adoption surges, but Australia's roadmap emphasizes safety institutes and voluntary guidelines. Newsrooms may thrive by specializing in human-centric reporting AI can't replicate. Optimistically, policy evolution could embed fairness, preserving Aussie voices.

Future of AI in Australian news media illustration

What This Means for Higher Education and Careers in Media Research

Universities like Sydney lead in probing AI-media intersections, training future experts. Programs in digital comms and journalism ethics boom. Aspiring profs and researchers, check opportunities to contribute.Professor jobs in media studies. Explore Rate My Professor for insights, higher ed jobs, and career advice.

Discussion

0 comments from the academic community

Sort by:
You

Please keep comments respectful and on-topic.

DLW

Dr. Liam Whitaker

Contributing writer for AcademicJobs, specializing in higher education trends, faculty development, and academic career guidance. Passionate about advancing excellence in teaching and research.

Frequently Asked Questions

🔍What are the main findings of the University of Sydney AI study on Australian news?

The study found Microsoft Copilot favors US/European sources in 80%+ of summaries, with Australian media in only ~20%. No local sources in 3/7 prompts, erasing journalists and regions.71

📊How was the methodology of Dr. Koskie's Copilot analysis?

434 summaries from 7 prompts over 31 days on Australian-configured Copilot. Focused on source bias, not accuracy.

🌍Why does AI sideline Australian journalism?

Replicates web power imbalances: dominant global sites prioritized, locals marginalized despite geo-tagging.

💸What impacts does this have on Australian media revenue?

Users read summaries without clicking through, cutting traffic and ad revenue, worsening news deserts.

🚀How is AI being adopted positively in Aussie newsrooms?

News Corp's NewsGPT automates routine tasks; boosts productivity for investigative journalism.

⚖️What policy changes does the study recommend?

Extend News Media Bargaining Code to AI; mandate geo-coding for local priority.

🏛️Are there risks to democracy from AI news summaries?

Yes: erodes trust in local news, homogenizes discourse, creates news deserts affecting civic engagement.

📈What does Reuters say about AI in Australian news?

2025 Digital News Report notes AI for news consumption rising; concerns over misleading content.

🎓How can researchers contribute to AI-media studies?

Join uni programs probing ethics. Check research jobs and career advice.

🔮What's the future outlook for AI in Australian journalism?

Balanced regulation could harness efficiency while protecting locals; watch 2026 AI Safety Institute.

📰Does the study address journalist anonymity in AI outputs?

Yes: Reporters homogenized as 'experts'; human labor erased, undermining accountability.