Since early 2025, the Indonesian Cyber Media Association (AMSI) has recognised that Indonesia’s digital media landscape is entering another major disruption, this time driven by artificial intelligence. AI-powered search and conversational platforms are changing how audiences find and consume information. Instead of opening links and visiting publishers’ websites, users increasingly receive summaries, explanations, and direct answers from Google AI Overview, ChatGPT, Gemini, Perplexity, and similar services.
Across AMSI’s membership, publishers have reported severe declines in organic traffic, with some losing more than 60%. Research conducted with Monash University Indonesia confirmed that this was not simply anecdotal. Among 44 Indonesian news outlets studied between September 2024 and November 2025, 37 recorded declining visitor numbers. The median decline was approximately 45%. Several groups of major and mid-sized publishers lost between 40% and 59% of visitors, while interviews with publishers identified traffic losses ranging from 20 to 70%
The study also found that Google AI Overview is increasingly acting as a gatekeeper between audiences and news websites. Users may still see publisher links, but they often stop after reading the AI-generated answer. For large publishers, the research identified a statistically significant decline in click-through rates after the expansion of AI Overview. Page views and organic search traffic were also settling at a lower baseline. This matters because many Indonesian publishers still depend on page views to generate programmatic advertising revenue.
Publishers face costs on both sides. They lose audience traffic while AI crawlers consume their content and server capacity. One publisher reported spending more than Rp100 million in additional infrastructure costs to prevent its website from slowing down or crashing. As one participant put it, publishers are paying for the infrastructure required to serve bots, while the economic value is captured elsewhere.
AMSI therefore commissioned a second study with PR2Media, a local media think tank, to examine whether its members were prepared to respond. The AI Readiness Index surveyed 262 media companies across Indonesia. The results showed a clear capacity gap. Large media companies scored 2.81 and were classified as “AI Aware.” Medium-to-large, medium, and small publishers scored 2.37, 2.35, and 1.98, respectively, placing them at the “AI Unaware” level.
The most serious weaknesses were not in willingness to experiment, but in governance, business strategy, data, and content protection. Almost half of the surveyed companies did not know how to protect their websites from AI crawlers. More than half were unfamiliar with frameworks that could provide controlled access to content for AI companies. Only 3.4% had implemented a combination of tools such as robots.txt, login authentication, Cloudflare, or a controlled scraping framework. Nearly half had not identified potential commercial partnerships with AI platforms.
“These findings made one point clear: individual publishers cannot solve this problem alone. This is especially true because more than 70% of AMSI members are small local media organisations outside Jakarta, often operating without dedicated technology, product, legal, or data teams”, highlighted Wahyu Dhyatmika.
With support from the International Fund for Public Interest Media, AMSI therefore began developing a locally governed AI and data-access solution using OpenMined. The objective is not to close journalism off from AI, but to replace uncontrolled scraping with permission-based access. Under this model, publisher data remains in a controlled environment. AI companies can submit queries, but publishers determine which content may be accessed, by whom, for what purpose, and at what price. Usage can be recorded, attribution preserved, and payments linked to actual access.
OpenMined’s Syft infrastructure provides the technical foundation. Syft Space acts as the controlled home for publisher data, access policies, and pricing. Syft Hub provides a directory through which AI companies, developers, enterprises, and researchers can discover available sources. Through an AMSI collective endpoint, members can participate without having to build separate infrastructure. Each publisher can retain its own attribution, analytics, control, and share of revenue.
“Our participation in the Journalism Cloud Alliance reinforced this approach. JCA demonstrates that shared technology, pooled expertise, common standards, and collective negotiation can give independent media access to infrastructure they could not afford or manage individually”.
AMSI is now testing the system with 50 publishers. The pilot covers CMS integration, data ingestion, AI queries, source attribution, access policies, and usage-based payments. By the third quarter of 2026, their data hub AMSInews.id aims to begin serving authorised AI platforms and developers through this collective infrastructure.
“The broader lesson is that sustainable media innovation requires more than teaching journalists to use AI. It requires collective ownership of infrastructure, stronger data governance, transparent attribution, enforceable access rules, and fair mechanisms for sharing value. Collective technology is becoming essential to ensure that local journalism remains visible, viable, and capable of serving its communities in the AI economy.”