Use of artificial intelligence at the UB

The CRAI supports the UB community in the critical and responsible use of AI, particularly in searching and evaluating information, citation, academic integrity, and access to specialised resources.

Key concepts

Artificial intelligence (AI)

Artificial intelligence (AI) is a field of computer sciene that develops systems capable of carrying out tasks that normally require human intelligence, such as learning, reasoning, recognizing patterns, interpreting information, planning or making decisions.

AI is a broad concept that encompasses different technologies and applications. One of these is generative artificial intelligence (GenAI), which is designed to create new content based on the data and patterns on which it has been trained.

Generative artificial intelligence (GenAI)

Generative artificial intelligence (GenAI) is a technology capable of generating content such as text, images, code, summaries, audio or video. It can be a useful tool for studying, answering questions, organising information or generating ideas. However, GenAI tools may invent data or generate incorrect, biased or unverifiable information.

More information:

• La IA i la IAG: dos conceptes que cal distingir (CRAI Teaching Unit)
• IAG: conceptes del procés d'aprenentatge (CRAI Teaching Unit)

Declaring, citing and referencing

In the academic use of artificial intelligence, three related but different practices should be distinguished:

• Declaring the use of AI: transparently explaining how an AI tool has contributed to the process of elaboration of a paper, research, or teaching material. The declaration may indicate which tool was used, for what purpose and at what stage of the work.
• Citing AI-generated content: attributing a specific answer, text, image, piece of code or other output generated by an AI tool, when appropriate and in accordance with the applicable citation style or guidelines. 
• Citing and referencing original sources: identifying the publications, data, images and other sources on which the information incorporated into the work is based.

Declaring the use of AI does not replace citing or referencing the original sources.

It's not always necessary to cite an AI tool simply because it has been used: it depends on the type of use, the content incorporated and the applicable academic, teaching or publishing regulations.

Institutional recommendation and good practices

UB recommends: Copilot

Copilot (Bing Chat Enterprise) is the GenAI tool institutionally adopted by the UB because it provides:

• A secure and protected environment to work with academic information.
• Guaranteed privacy: the entered data is not used to train models.

Copilot on the ATIC Software Services UB website Guia d'ús de Copilot (Bing Chat Enterprise) 

Good practices for the use of GenAI at UB

Before using any GenAI tool, it is essential to know and apply the Best practices for the use of generative artificial intelligence at the University of Barcelona. These guidelines explain how to use these tools while respecting academic integrity, data security and information quality.

   Best practices for the use of GenAI at the UB

General recommendations for responsible use

AI tools can be useful for exploring information, summarising content, generating ideas or supporting various academic, teaching, and research tasks. However, they do not replace critical reading, personal judgement, or consultation of original sources.

Before incorporating AI-generated results into an academic paper, research project or report, the following recommendations should be taken into account:

• Verify data, authors, dates, and references.
• Avoid introducing personal, confidential or sensitive data if the tool's terms of use do not guarantee appropriate handling.
• Properly cite the sources used.
• Review, modify and validate generated content before incorporating it into the final result.
• Follow academic regulations, teaching staff instructions and applicable institutional or publishing policies.

The author is always responsible for the content presented, including errors, omissions, incorrect references, and potential biases arising from the use of AI tools.

The ethical risks associated with the operation of these tools and the results they generate must also be considered, as they may affect academic activity and the way we access, use and produce information. These risks include the generation of fake or biased information, the reproduction of discrimination, lack of transparency, impacts on privacy and conflicts related to intellectual property.

Furthermore, the generative artificial intelligence models have an environmental impact resulting from the energy and water consumption of data centres, as well as the carbon footprint associated with training and the large-scale use of these models.

Related resources:

• La IA al CRAI: ús responsable. Infographic by the CRAI UB
• Desenvolupar l’actitud ètica en l’ús de la IA generativa. Portal TAC TIC

How to declare the use of AI

Declaring the use of artificial intelligence means transparently explaining how an AI tool has intervened in the development of an academic paper, teaching material, research project, or publication.

The use of these tools must be declared when required by academic regulations, teaching staff instructions, or institutional, publishing, journal, conference, or funding body policies. It is also recommended to declare their use when the tool has played a significant role in developing the work.

The declaration should include, at minimum:

• The tool and, whenever possible, the version used.
• The purpose for which it was used.
• The stage or parts of the work in which it was used.
• The type of assistance provided.
• The review, verification, or modification carried out by the author.

Example of declaration:

In preparing this work, [name of the tool and version] was used for [purpose or task]. The generated content was reviewed, cross-checked, and, where necessary, modified by the author, who assumes responsibility for the final result.

For a more detailed declaration, GAIDeT (Generative AI Delegation Taxonomy) can be consulted. This taxonomy includes an online generator that allows users to describe, in a structured way, the tasks delegated to AI tools during the different stages of academic and scientific work.

Declaring the use of AI does not legitimise every use of these tools and does not replace citing original sources or including the corresponding bibliographic references.

AI according to user profile

AI for students

The use of generative artificial intelligence tools in academic work must be permitted by the teaching staff and must comply with the course syllabus, course guidelines and applicable academic or assessment regulations.

Before using a GenAI tool, students should check:

• Whether its use is permitted
• Which activities or stages of the work it may be used for.
• Which uses are not acceptable.
• Whether its use must be declared.
• Whether any generated content must be cited.

In case of doubt, students should consult the responsible professor beforehand.

Declaring the use of AI is an exercise in transparency that helps preserve responsibility, honesty and academic integrity. It does not legitimise every use of AI, nor does it replace consulting and citing academic sources.

The CRAI has produced three videos to guide students on how to declare and, where appropriate, cite the use of GenAI in academic work:

• Has utilitzat la IA? Ho has declarar i, quan correspongui, citar!
• Has utilitzat la IA? Ho has d'explicar!
• Has utilitzat la IA? Ho has de verificar i contrastar!

These materials are framed within the Best practices for the use of generative artificial intelligence at the University of Barcelona and the Code of ethics on integrity and best practices at the University of Barcelona.

Related resources:

• How to cite AI use in APA style
• User training in GenAI for students (Office of the Vice-Rector for Digitalization Policy). Access to online courses (certified and non-certified courses)

AI for researchers

In the field of research and scholarly publishing, GenAI tools cannot be considered authors, and their use must be transparently disclosed in accordance with the policies of each journal, publisher, conference, institution, or funding body.

According to the recommendations from organisations like the Committee on Publication Ethics (COPE) and leading publishers (Elsevier, Springer, Wiley,...), AI cannot assume legal or ethical responsibilities, and responsibility for the generated content always lies with the human authors.

When GenAI tools are used for writing, translating, language editing, summarisation, data analysis, code generation or other stages in the research process, the relevant publishing policy must be followed. The declaration can be included in the methodology section, acknowledgements, or the specific section provided by the journal or publisher.

To describe the use of AI more systematically across the different stages of academic or scientific work, models and tools such as GAIDeT can be consulted. This taxonomy includes an online generator that makes it possible to identify and disclose the tasks delegated to GenAI during the different stages of research and scholarly publishing.

Related resources:

• User training in GenAI for PDI (Office of the Vice-Rector for Digitalization Policy). Access to online courses (certified and non-certified courses)

AI for teaching staff

This section focuses particularly in generative artificial intelligence (GenAI), one of the AI-related technologies having the greatest impact on the planning, design and support for teaching activities.

GenAI tools can provide support for teaching staff at different stages of teaching:

• course planning
• design of learning activities
• preparation of teaching materials
• development of rubrics
• review or adaptation of content

Their use must always be guided by teaching expertise and the learning goals of each course. GenAI doesn't substitute the responsibility of teaching staff in the design, monitoring, and evaluation of educational activities. Rather, it can act as a support tool for generating ideas, exploring alternatives, or improving specific teaching processes.

To obtain more useful responses, it is important to formulate good prompts: provide sufficient context, specify the task, indicate the expected format and, where appropriate, provide examples.

When AI has played a significant part in the preparing teaching materials or other resources, it is recommended to consider whether its use should be declared following the general recommendations on transparency.

Teaching staff are also advised to make it clear to the students:

• whether the use of GenAI tools is permitted
• under what conditions
• at which stages of the work
• when its use must be disclosed
• when generated content must be cited
• which uses are not acceptable
• how the activity will be assessed

Related resources:

• Planificar la docència amb IAG (CRAI Teaching Unit)
• Com fer bons prompts per la IAG (CRAI Teaching Unit)
• Competència docent en IAG (Institut de Desenvolupament Professional)
• User training in GenAI for PDI (Office of the Vice-Rector for Digitalization Policy). Access to online courses (certified and non-certified courses)

Practical resources and guidance

Prompts

A prompt is an instruction, question, or text entered into a generative artificial intelligence tool to indicate the type of response or content that is required.

To obtain more useful results, it is recommended to formulate clear, specific and contextualized prompts. A good prompt may include:

• the role or perspective the tool should adopt
• the task it should perform
• the context or background information
• the expected format of the response

When using GenAI tools, always remember to review the results, cross-check the information against reliable sources, avoid entering personal, confidential or sensitive data, and disclose their use when relevant, in accordance with the University of Barcelona's good practice guidelines.

Related resources:

• Recommendations and examples provided by Copilot (GenAI tool institutionally adopted by the UB)
• Com fer bons prompts per la IAG (CRAI Teaching Unit)
• Domina l’art dels prompts: 10 tips essencials per generar prompts efectius (Institut de Formació Contínua-IL3)

GenAI applications

Currently, there are many GenAI applications that can create, transform or analyse content in different formats, including text, images, audio, video, code, presentations, diagrams, summaries, and teaching materials, among others.

Before using any tool, it is important to review its terms of use, privacy options, the way the data entered is processed, and the possible limitations of the generated results. It should also be kept in mind that responses may contain errors, bias, or unverifiable information, and must always be critically reviewed.

To consult a selection of GenAI tools and applications organised by use, you can access the guia elaborada per la Bibliotècnica de la UPC*.

*The inclusion of a tool in this selection does not imply institutional recommendations nor validation of its privacy conditions. Before using it, users should review its terms of use, data policy, and limitations.

Resources subscribed by the CRAI with AI functionalities

The CRAI provides the UB community with subscribed resources that incorporate AI functionalities. These tools can support information searching, bibliographic explorations, information analysis, and the generation of structured answers based on academic or specialised sources.

It should be noted that AI-generated results must be critically reviewed, cross-checked against the original sources and appropriately cited according the academic or professional context.

Resource	What is it?	What is it used for?	Access
Web of Science Research Assistant	A GenAI module integrated into Web of Science. It allows users to make natural-language queries, obtain overviews of results, explore research topics, identify research trends, and use interactive visualisations based on Web of Science Core Collection data. The CRAI of the UB has subscribed this module as a complement to its Web of Science subscription and makes it available to the entire UB community through Cercabib. More information: Seminari web sobre l'assistent IA de Web of Science al canal de YouTube del CRAI de la UB.	It is useful for guiding a bibliographic search, exploring a new topic, identifying related concepts and obtaining an initial overview of a research area. For literature reviews, it should only be used as an exploratory support tool, and its results must always be cross-checked through verifiable documentary searches.  Clarivate describes the tool as an assistant specifically designed for academic research, with multilingual support, task-oriented guidance and interactive visualisations.	Cercabib > Web of Science. Subscription valid until December 2026.
PharmaPendium AI	An AI tool published by Elsevier for searching and analysing regulatory information in the pharmaceutical sector and for drug development. It integrates a large amount of preclinical and clinical trial data, including toxicity, adverse events, and side effects. It also allows users to identify drugs related to specific indications through searches based on diseases, symptoms, adverse events and other biomedical criteria.	Allows natural-language queries of regulatory data and documents from the Food and Drug Administration (FDA) and the European Medicines Agency (EMA). The answers are based on PharmaPendium content.  It can be particularly useful for users in pharmacy and biomedicine.  Results must always be verified against the original regulatory documents and do not replace expert judgement or professional advice in regulatory, clinical or pharmaceutical development processes.	Cercabib > PharmaPendium IA
Reaxys	An Elsevier chemical database that, since 2025, has incorporated the Predictive Retrosynthesis module, featuring AI developed by Iktos and trained  on Reaxys reaction data.	Proposes synthesis routes (retrosynthesis) taking into account stereochemistry, cost and reagent availability. Routes must always be validated through expert judgement and the relevant literature.	Cercabib > Reaxys
Statista	A statistical data and market research platform that incorporates Research AI, an assistant for natural-language queries.	Based on a question, it selects the most relevant Statista sources and generates a summarised and cited response. It is useful for an initial approach to market or consumer data, but figures should always be cross-checked against the source.	Cercabib > Statista
Aranzadi Infinita	A legal database from Aranzadi La Ley with several AI modules.	One of the modules is Jurimetría, which analyses large volumes of court decisions to identify trends (such as the duration of proceedings and lines of case law) using interactive graphs. It does not replace professional judgement or direct consultation of court decisions.	Cercabib > Aranzadi Infinita
JSTOR	A database of articles and books in the humanities and social sciences, with an AI tool to support research.	Allows semantic search in natural language and, for enabled documents, summarises key points and answers questons about the text. Users must register beforehand to activate it; summaries should always be checked against the original.	Cercabib > JSTOR
PsycArticles	A psychology database from APA (PsycNet platform) featuring a Natural Language Search based on semantic search.	Allows users to formulate searches as natural-language questions rather than keywords.  It is useful for exploring a topic when the appropriate search terms are not yet clear, complementing the search with descriptors when comprehensiveness is required.	Cercabib > PsycArticles

Recommended reading
 

Com fer un ús responsable de la intel·ligència artificial en educació?

Com hem de citar la IA en els treballs?

Guía para el uso de IA generativa en educación e investigación

Intel·ligència artificial en la docència

Using generative artificial intelligence to produce texts in English

More institutional AI resources

Artificial intelligence at the Dipòsit Digital de la UB Artificial intelligence at the TÀCTIC blog  Artificial intelligence at the Office of the Vice-Rector for Digitalization Policy Documents about artificial intelligence made by the CRAI (Dipòsit Digital) 

Last updated: September 2026