Artificial intelligence (AI) technology, which is considered one of the most important stages of technological development, has become popular in recent years and has triggered important discussions about how different professions will be affected by this technology. Journalism, where speed is extremely important, is one of these professions. Based on this, the research aims to compare the news production methods of the two popular AI applications, ChatGPT and DeepSeek, from a social sciences perspective. To this end, both AI models were instructed to generate three news articles on specific topics, the resulting texts were analysed using document analysis techniques, and the findings were presented using descriptive analysis techniques. According to the findings, both artificial intelligence applications generally adhere to the 5W 1H principle (who, what, when, where, why, and how) in the creation of news content. ChatGPT is careful and transparent in citing its sources, while DeepSeek presents events in a multifaceted manner. However, inconsistencies were identified in some factual details across the texts produced, and neither application fully conformed to the linguistic and orthographic standards set by the Turkish Language Association (TDK). The findings of the study reveal the extent to which prominent artificial intelligence models can uphold the fundamental values of journalism in the news production process, indicating that they should be subject to scrutiny in terms of source diversity and transparency, ethical issues, and linguistic accuracy. The study recommends that AI-generated news content be subject to structured editorial oversight, include clear labeling of AI involvement, and that media professionals receive training in the effective and ethical use of AI tools. This discussion aims to contribute to multidimensional debates on artificial intelligence, journalism, and access to accurate information.
Teknolojik gelişmelerin ulaştığı en önemli aşamalardan biri olarak kabul edilen yapay zekâ teknolojisi son yıllarda popüler hale gelmesiyle birlikte meslek gruplarının bu teknolojiden hangi biçimlerde etkileneceğine yönelik önemli tartışmaları tetiklemiştir. Bu meslek grupları arasında hızın öneminin son derece görünür olduğu gazetecilik de yer almaktadır. Buradan hareketle araştırma, popüler yapay zekâ uygulamaları olan ChatGPT ve DeepSeek uygulamalarının haber üretim biçimlerinin sosyal bilimler perspektifinden karşılaştırılmasını amaçlamaktadır. Bu doğrultuda her iki yapay zekâ modeline belirli konularda üçer haber yazması komutu verilmiş, elde edilen metinler doküman analizi tekniğiyle analiz edilmiş, bulgular ise betimsel analiz tekniğiyle sunulmuştur. Elde edilen bulgulara göre, her iki yapay zekâ uygulaması da haber metinlerinin oluşturulmasında 5N 1K kuralına genel anlamda sadık kalmaktadır. ChatGPT başvurduğu kaynakları belirtme konusunda özenli ve şeffaf davranırken; DeepSeek olayları çok yönlü biçimde ele almaktadır. Verilen bazı bilgiler arasında çelişki olduğu gözlemlenen uygulamaların yazım ve imla kuralları söz konusu olduğunda Türk Dil Kurumunun (TDK) standartlarına tam anlamıyla uymadıkları gözlemlenmiştir. Araştırmanın bulguları, öne çıkan yapay zekâ modellerinin haber üretimi aşamasında gazeteciliğin temel değerlerini ne ölçüde sürdürebildiğini ortaya koyarak kaynak çokluğu ve şeffaflığı, etik hususlar ve dilbilimsel açılardan denetime tabi tutulması gerektiğine işaret etmektedir. Uygulamaların bu hususlar göz önünde bulundurularak yeniden yapılandırılması, yapay zekâ modelleri ile üretilen haberlerde etiketleme sistemine bağlı kalınması ve medya çalışanlarının yapay zekâ kullanımı konusunda eğitilmesi gibi öneriler sunulmaktadır. Bu tartışma ile yapay zekâ, gazetecilik ve doğru bilgiye erişime dayalı çok boyutlu tartışmalara katkı sunulması amaçlanmaktadır.
Primary Language | Turkish |
---|---|
Subjects | Journalism Studies |
Journal Section | Articles |
Authors | |
Publication Date | July 20, 2025 |
Submission Date | April 28, 2025 |
Acceptance Date | July 13, 2025 |
Published in Issue | Year 2025 Volume: 34 Issue: Uygarlığın Dönüşümü - Sosyal Bilimlerin Bakışıyla Yapay Zekâ |