Who is writing the news? Artificial Intelligence, journalism in the US and Turkey

It is generally believed that the structural and technological changes in the media industry and the solutions developed against them in the US usually manifest in Turkey 10 years later. At the very least, we are currently seeing the reflections of this in the search for digital media business and revenue models. For this reason, it is important to evaluate the data from two studies conducted by the Pew Research Center in the US in 2024 and 2025 regarding the impact of artificial intelligence (AI) on news production and the journalism profession over the next 20 years. The general assessment of the research is that Americans have a largely negative attitude toward AI-written news. Looking at the details, according to data from 900 US adults who agreed to share their online browsing activities, 50% of participants think that AI will have a negative impact on news. 59% say that AI will create fewer jobs for journalists. This pessimism becomes more pronounced as the level of education increases. University graduates have a more negative perspective than high school graduates.

Evaluating AI's current news-writing skills, 41% of Americans think that AI writes worse news than journalists, 19% think AI writes better, and 20% think that AI and journalists produce work of similar quality. These results show that trust in artificial intelligence is limited, but its potential is not ignored.

66% of Americans are seriously concerned about the potential for artificial intelligence to spread misinformation. This rate is almost equal among both Republicans and Democrats. The results show that there is a non-partisan distrust of artificial intelligence.

GOOGLE SEARCH ENGINE SUMMARIZES NEWS WITH AI

Google's new "AI Overview" feature will significantly change the way news and information are accessed. 58% of users have encountered an AI summary at least once. In search results with AI summaries, the click-through rate for users is 8%, whereas this rate is 15% for results without summaries. These data show that artificial intelligence is shifting content consumption in favor of platforms and to the detriment of publishers. Looking at the dynamics of AI summary production, searches that are longer and contain question sentences increase the likelihood of generating an AI summary. For example, 53% of queries with 10+ words generate an AI summary. 60% of queries starting with question words like "what, why, who" result in an AI summary. This use of AI in news writing shows that it will diminish the human-specific aspects of journalism, such as creative and in-depth analysis, and consequently, the quality of the public's access to news.

MONOPOLIZATION OF INFORMATION WILL INCREASE WITH GOOGLE AND AI SUMMARIES

The forecast for a shrinking workforce in journalism over the next 20 years is an indicator that structural transformation and crises in the media industry will increase. According to the research, the most frequently cited sources in both Google AI summaries and standard search results are: Wikipedia, YouTube, and Reddit. This situation should be interpreted as an indicator that Google will increase the monopolization of information with its AI summaries. The decrease in users' need to click on news and publisher sites will harm the traffic and revenue of news platforms, and therefore the sustainability of journalism. The use of AI by search engines will make societies dependent on information filtered only by technology companies. Although the US public currently does not find AI-written news to be of high quality (41%), the capacity of AI technologies to improve themselves at great speed—through human-machine communication and computer-mediated communication—brings up the potential for AI content that has not passed through an editorial filter to create information pollution and manipulation in society.

WHAT MIGHT HAPPEN WHEN AI SUMMARIES IN SEARCH ENGINES START IN TURKEY?

Journalism in Turkey is already in a difficult situation due to press freedom restrictions, economic pressures, censorship, self-censorship, and political polarization. In such an environment, the impact of artificial intelligence could be much deeper and riskier than in the US. AI-supported news production could further weaken editorial independence. For example, in the current political environment, if algorithms are trained according to state policies, AI-supported production of "official discourse-compliant" news could become a tool to suppress diverse and opposition voices.

Some news sites in Turkey are already known for clickbait content and low-quality news. When AI content is used without oversight, the risk of misinformation, lack of context, or "hallucinations" increases. This will further weaken the quality of public information.

A significant portion of media workers in Turkey work without unions, as freelancers, temporarily, or for low wages due to the surplus of reserve labor. Automation will further increase job losses, especially in local news networks and small media organizations. Editing texts from agencies with AI could lead to journalism being done not in the field, but through desk-based simulations.

Digital media news organizations in Turkey access advertising revenue largely through Google and social media traffic. If AI summaries with "Google AI Overview" become widespread in Turkey, users will click on news sites less, and the already limited revenue streams for independent digital media organizations will decrease even further.

If AI-supported content in Turkey is combined with fake news and propaganda on social media, public opinion could be steered much faster. Especially during election periods or moments of crisis (earthquakes, conflicts, fires, economic collapse), AI-supported information manipulation will have negative social impacts.

A large-scale public opinion survey similar to Pew regarding AI news production and writing has not yet been conducted in Turkey. However, as a hypothesis, since media trust is low, readers and users may be skeptical about AI producing news. According to the Reuters Institute's 2024 Digital News Report, trust in news sources in Turkey is only 36%. For this reason, an illusion that artificial intelligence will produce "more neutral" news may also form. Especially among digital natives and young people, it is possible that summary news coming via AI will be perceived positively in terms of "getting information quickly."

In Turkey, where freedom of expression is limited, media revenue is weak, and AI literacy is low, the transformation in accessing news through search engines and AI summaries should be evaluated together with regulations and legislation based on the public's right to receive news rather than media capital, proactive media policies, public education, and ethical structures. Media institutions and society in Turkey must know and question the sources from which AI summaries are created. In this context, it should become a legal obligation to clearly label every piece of content produced by AI, ethical guidelines should be established, and content control mechanisms should be set up to combat disinformation, especially during election periods.