Use of AI in research, writing and publications: DeepSeek vs. ChatGPT

Abstract

The integration of artificial intelligence (AI) into research writing and publications has transformed scholarly communication, with models like ChatGPT and DeepSeek leading the charge. While ChatGPT, released in 2022, gained widespread adoption for its fluency and versatility, DeepSeek emerged in 2025 as a formidable open-source alternative, excelling in technical writing, citation accuracy, and computational efficiency. DeepSeek’s strengths lie in structured academic sections (e.g., methods, results), minimal citation hallucinations, and offline usability, making it ideal for STEM domains. In contrast, ChatGPT excels in narrative-driven content but struggles with precision and fabricates references more often. Cost and accessibility further differentiate the two: DeepSeek is fully free, modifiable, and self-hostable, whereas ChatGPT’s advanced features require subscriptions and rely on cloud infrastructure. Ethical transparency remains a challenge for both.

The future of AI in research demands rigorous human oversight, with researchers cross-verifying outputs and journals enforcing transparency in AI use. As AI tools evolve, their success will hinge on balancing technological capabilities with ethical accountability, ensuring integrity in scholarly work. DeepSeek’s open-source model and technical prowess position it as a preferred choice for academia, though ChatGPT retains an edge in creative and multilingual tasks. Ultimately, AI’s role in research hinges on responsible adoption and continuous refinement by the scholarly community.