U.S. Local Media: 1 in 10 Articles Written by AI
University of Maryland Study Reveals Over 1,500 Newspapers Using AI Automation Tools… Survival Strategy or Journalism Crisis?

- •Approximately 9% of articles from over 1,500 U.S. newspapers are written by AI, with some exceeding 20%.
- •AI adoption is spreading as a solution to staff and budget shortages caused by 3,500 local newspaper closures over the past 15 years.
- •Cases of false information publication due to AI 'hallucination' have emerged, making transparency and fact-checking enhancement key challenges.
Silent Spread of AI Journalism
According to research from the University of Maryland, approximately 9% of articles published by over 1,500 U.S. local newspapers are written by artificial intelligence (AI). For some newspaper groups (Boone Newsmedia), this ratio has been confirmed to exceed 20%.
Since the emergence of large language models (LLMs) like ChatGPT, U.S. local news organizations have been utilizing AI tools for routine article writing such as city council meeting minutes, local event coverage, and rewriting press releases. This is a quiet transformation happening largely unbeknownst to readers.
A Choice for Survival
The background to local media's dependence on AI lies in the structural crisis of the American newspaper industry. Over 3,500 local newspapers have closed in the past 15 years, and surviving news organizations face challenges of limited budgets and declining readership.
AI tools can generate articles that couldn't be covered due to staff shortages for a monthly cost of just tens of dollars. For many journalists, this is accepted not merely as a convenience but as a survival strategy. By having AI write local news that limited staff couldn't cover, it can partially fill the information void in local communities.
Credibility Issues and 'Hallucination' Risks
However, AI journalism carries serious ethical and technical problems. Generative AI is notorious for the 'hallucination' phenomenon of fabricating or distorting facts. In fact, in 2025, the Chicago Sun-Times published a summer reading list written by AI, where some book titles and author names turned out to be completely fictitious.
AI also tends to produce unnatural writing styles, repetitive expressions, and hackneyed clichés. Editors face the burden of manually verifying and correcting all AI-generated content. If this process is neglected, the news organization's credibility can be rapidly undermined.
Efforts to Ensure Transparency
To address these issues, new codes of practice are emerging in the American journalism industry.
- Disclosing AI authorship: Clearly indicating whether AI tools were used in articles
- Journalist training: Conducting training to identify AI-specific writing styles and repetitive patterns
- Strengthening fact-checking: Reinforcing fact-verification procedures for AI-generated content
Some news organizations are choosing to use AI as an assistive tool rather than full automation. This is a hybrid model where AI handles draft writing and data organization, but human journalists must perform final verification and editing.
France's Cautious Approach
Unlike the United States, the French journalism community is taking a conservative stance on AI adoption. In 2025, the French Journalists' Union (Syndicat national des journalistes) recommended "prohibiting the publication of articles written entirely by AI" and proposed mandatory explicit labeling when used.
While some French news organizations are exploring AI in controlled experimental forms, large-scale adoption remains distant. European journalism's slower AI adoption is attributed to a cultural background that prioritizes journalistic ethics and quality standards.
[AI Analysis] Journalism at a Crossroads
U.S. local media's dependence on AI demonstrates the tension between short-term survival and long-term credibility. AI is cost-effective and fast, but has fundamental limitations of errors, bias, and lack of creativity.
AI journalism is likely to evolve in two directions going forward.
1. Strengthening Regulation and Transparency: Like the European model, restricting AI use and mandating disclosure
2. Establishing Hybrid Models: AI handles simple tasks while humans provide analysis and context
With more than half of U.S. local news organizations expected to expand AI adoption within the next three years, the key will be not the speed of technology adoption, but whether journalistic principles are upheld. Media that lose reader trust cannot survive.
Korean journalism should also monitor these changes. Finding balance between AI tool efficiency and journalistic ethics will be a challenge all news organizations will face going forward.
댓글 (3)
기사 잘 봤습니다. 다른 시각의 분석도 읽어보고 싶네요.
Media:에 대해 더 알고 싶어졌습니다. 후속 기사 부탁드립니다.
그 부분은 저도 궁금했습니다.
More in AI & Tech

영국 정치지도자들, 아동 성착취 혐의에 대한 긴급 조사 촉구

A humanoid robot performing in China has a child's face on it.

U.S. jury finds Meta and Google responsible for ‘social media addiction’… 3.7 billion won compensation ruling

Japanese X-ray Observatory makes first direct measurement of ultrafast 'cosmic wind' in galaxy M82

NASA selects 24 people for 2026 Astrophysics Postdoctoral Fellowships

Ethereum is at a crossroads to ‘redefine its identity’ ahead of the quantum computing and AI era
Latest News

"간부 잘 아는데 교통비 좀" 휴가 군인들 돈 뜯은 50대 구속
50대 A씨가 휴가 중인 군인들에게 부대 간부를 아는 척 접근해 돈을 사취

英 옥토퍼스, 이란 전쟁 이후 태양광 판매 50% 증가
이란 전쟁 이후 영국 옥토퍼스의 태양광 판매량 50% 증가

당정 "추경, 지방·취약계층에 더 지원되는 방식으로"
당정이 지방자치단체와 취약계층 중심의 추경 편성 방침 재확인

당정, 석유 최고가격제 손실 보전을 추경에 반영키로
당정이 석유 최고가격제 손실을 추경에 반영하기로 결정

어머니 폭행하고 금팔찌 빼앗은 30대 아들 경찰에 붙잡혀
어머니 폭행 후 금팔찌 빼앗은 30대 남성 체포

아이티 갱단 폭력사태로 10개월간 5천명 이상 사망
아이티에서 지난 10개월간 갱단 폭력으로 5천명 이상 사망

서방 정보당국 "러시아, 우크라이나 전쟁 후 이란에 드론·식량 공급"
서방 정보당국, 러시아의 이란 드론·식량 공급 작업 거의 완료 파악

6년 전 세 살 딸 살해한 30대 친모 구속송치
경찰, 6년 전 세 살 딸 살해 혐의 30대 친모를 구속송치