On February 1, 2026, the following came into effect in Ukraine: Law No. 4742-IX “On Academic Integrity.” For the first time, a separate concept has appeared in the legislation:
"Unethical use of artificial intelligence"—on par with plagiarism and data fabrication.
There's also a Registry of Violators. And it's entirely possible to end up on it. Most researchers don't know that their text is already undergoing automatic AI test at universities.
Before analyzing the errors, it's important to understand the context: What exactly did the law on academic integrity and AI change in 2026? — We discussed this in detail in a separate article.
Test yourself: Are you making any of these 5 mistakes?
Mistake #1: Inserting AI-generated text without checking it or providing a source
The most common. And the easiest to detect.
Ukrainian universities have switched en masse to StrikePlagiarism with an AI-content detection module—it went live in December 2025. The system generates an official PDF report containing two metrics: the percentage of plagiarism and the percentage of text likely generated by a machine.
At the same time, Scopus journals (Elsevier, Springer Nature) require a separate declaration regarding the use of AI directly in the manuscript. Without it, the submission will be rejected.
If you want to do AI text verification Before serving, we recommend using special utensils. Top 5 Free AI Detectors for Academic Texts — A detailed review with a comparison of accuracy.
What to do: Please list the tools (ChatGPT, Claude, Gemini), specify the tasks for which you used them, and at what stage you used them. The law and Cabinet of Ministers of Ukraine Resolution No. 928 They demand this explicitly.
Mistake #2: Not verifying the sources provided by AI
This isn't just a mistake—it's a ticking time bomb.
In December 2025, it was revealed at the NeurIPS conference that 100+ Bizarre Quotes in papers by authors from Google, Harvard, Meta, and Cambridge. In another case—the University of Hong Kong—20 of the 61 citations in a peer-reviewed article turned out to be completely fabricated.
The statistics are disheartening: ChatGPT generates between 39.6% and 55% completely fabricated citations when creating literature reviews.
There's only one rule: Check every source obtained through AI manually—either in Google Scholar or directly in the database.
Mistake #3: You use AI to translate and don't mention it
Good news: In its guidelines, the Ministry of Education and Science explicitly permits the use of AI for translating and adapting quotations.
Bad: Concealing this fact is already a violation.
If you translated a foreign source using ChatGPT or DeepL and pasted it “as is” without citing it—that falls under the law. A single sentence in the methodology or “Acknowledgments” section completely resolves the issue.
Mistake #4: Letting AI Draw Scientific Conclusions for You
Cabinet of Ministers of Ukraine Resolution No. 928 and Law No. 4742-IX are equally clear: All parts of the text written using AI must be reviewed and edited by the author personally.
The Ministry of Education and Science specifically prohibits “the generation of scientific conclusions using AI without independent verification.”
In other words, even if the AI wrote a brilliant conclusion—it can’t just be left “as is.” This isn’t bureaucracy. It’s a safeguard to ensure that the conclusion isn’t based on fabricated data. Learn more about the legal requirements for the use of AI in academic work.
Mistake #5: You don't keep a paper trail of your AI usage
The registry of violators is required by law and will become operational after July 31, 2026, when each university is required to adopt internal regulations. The responsible authority is NAZYAVO.
The consequences of being included in the Registry are not merely symbolic:
• at least 1 year without the right to hold academic titles;
• revocation of an academic degree if discovered after the defense;
• a 10-year ban on holding leadership positions at higher education institutions in systemic cases;
• Withdrawal or Revocation of a Diploma For applicants.
A simple practice that offers protection: Keep an AI log from day one—date, tool, query, and purpose. And include the statement directly in the text of your work.
Declaration template under Law No. 4742-IX:
“The following artificial intelligence tools were used in this work: [names]. They were applied to [specific tasks]. All sections written with the use of AI have been reviewed and edited by the author personally. The scientific arguments, conclusions, and results represent the author’s own intellectual contribution.”
Conclusion
AI isn't banned—and that's the main thing. But 2026 changed the rules of the game forever.
Previously, the question of “whether I used AI” was a matter of personal conscience. Now it is a legal obligation enshrined in law, monitored by the university, and subject to very specific penalties for violations—including revocation of one’s degree.
Almost everyone makes the 5 mistakes listed in this article—not because they’re dishonest, but because the new requirements were introduced recently and haven’t yet become second nature. The registry of violators will go into effect after July 31, 2026. Until then, there’s still time to get things in order.
Three things you should do today:
• Start keeping an AI journal—tool, date, query, goal;
• Add a statement regarding the use of AI in your work;
• How to Run an AI Check — A Comparison of the Top 5 Free Tools — A comparison of the top 5 free tools.
This isn't bureaucracy. It's a safeguard for your degree.
Are you working on your dissertation and don't want to take any chances?
The Article Academy handles the entire process—from structuring and writing the sections to AI and plagiarism checks and support throughout the defense process. Official contract, guarantees, confidential.
