You are here:

Use cases of generative artificial intelligence in research staff’s scientific production

Generative Artificial Intelligence (GenAI) can support scientific production (literature, article sections, research project writing) – applications useful to organise content, develop ideas and personalise your learning experience in a conscious way.

The University of Bologna encourages a responsible and conscious use of these tools.

GenAI should enhance critical thinking and personal creativity without replacing them.

The following examples of use cases cover scientific production by presenting some of the most widespread applications.

Summarising scientific articles or other documents

Creating brief abstracts from long documents

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Risk that the abstract is inaccurate.
Risk of quoting other authors' opinions or thoughts inaccurately.
Risk of copyright infringement and confidentiality breach.

Suggestions and actions to complete

Check that abstracts are accurate and representative of the document content.
Make sure that key information is correctly summarised without neglecting crucial concepts or distorting the original meaning.
Check that other authors' opinions or thoughts are quoted in a complete and accurate manner and that the source is interpreted correctly.
Check that you have specific permission to upload to the GenAI system copyrighted or confidential documents, or documents subject to other contractual obligations.

Summarising the main conclusions and methodologies used in a scientific study

It complies with the University Policy principles.
It requires citing the use of GenAI.

Key points of attention

Risk that the summary is inaccurate.
Risk of quoting other authors' opinions or thoughts inaccurately.

Risk of copyright infringement and confidentiality breach.

Suggestions and actions to complete

Check that the summary of conclusions and methodologies is accurate.
Check that the key elements of the study are reported correctly, avoiding the omission of crucial details or distortion of results and techniques.
Make sure that other authors' opinions or thoughts are quoted in a complete and accurate manner and that the source is interpreted correctly.
Check that you have specific permission to upload to the GenAI system copyrighted or confidential documents, or documents subject to other contractual obligations.

Generating the outline or structure of an article from the user's draft

Generating an outline or table of contents from the author's draft (e.g. abstract, title)

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Risk of quoting other authors' opinions or thoughts inaccurately.

Suggestions and actions to complete

Make sure that other authors' opinions or thoughts are quoted in a complete and accurate manner and that the source is interpreted correctly when generating an outline or table of contents.

Creating the initial structure for the introductory section of a publication, then writing the text yourself based on that structure

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Risk of copyright infringement.

Suggestions and actions to complete

Make sure that you always quote other authors' thoughts when creating the initial structure for the introductory section of a publication.

Generating an article section

Generating a text from an outline of contents drafted by the user

It complies with the University Policy principles.
It does not require citing the use of GenAI, unless the generated text prevails over the researcher's own content.

Key points of attention

Risk of nonsense in the text.
Risk of copyright infringement.

Suggestions and actions to complete

Make sure that other authors' opinions or thoughts are quoted in a complete and accurate manner and that the source is interpreted correctly.
Check that the generated text makes sense.

Using non-conversational tools based on autocomplete

It complies with the University Policy principles.
It requires citing the use of GenAI, if the generated text contains elements of novelty.

Key points of attention

Risk of nonsense in the generated text.

Suggestions and actions to complete

Check that the generated text makes sense and is consistent with the rest of the content.

Generating a research project section from the user's draft

Using GenAI tools for drafting a project proposal

It complies with the University Policy principles.
It requires citing the use of GenAI.

Key points of attention

Risk of violating the GenAI acceptable use policy of the funding body or programme to which the project proposal is submitted.
Risk that the generated content and citations are inaccurate and inappropriate.
Risk of copyright infringement.
Risk of including out-of-scope activities.

Suggestions and actions to complete

Acceptable if it is accepted by the funding body and/or programme to which the project proposal is submitted, without prejudice to:

The proponent's full accountability for the generated content.
The obligation of disclosing the tools used.

Check that GenAI can be legitimately used pursuant to the rules of the programme, call for applications or guidelines published by the funding body.
State the GenAI tools you used in a transparent way.
Comply with privacy and intellectual property regulations and avoid actions that are unlawful or harmful to third parties (e.g. plagiarism).
Check the accuracy, validity and appropriateness of the generated content and citations, correcting any errors or inconsistencies.
Provide a list of sources used for generating content and citations, including those generated by a GenAI tool.
Check that citations are true, accurate and referenced correctly.

Generating a SOTA or related works section or chapter

Generating the entire state of the art section or chapter from GenAI alone

It complies with the University Policy principles.
It requires citing the use of GenAI.

Key points of attention

Risk that the generated information is incomplete because GenAI tends to find established knowledge and may overlook themes at the forefront of research.
Risk of wrong information.
Risk of copyright infringement.

Suggestions and actions to complete

Using GenAI is not recommended from a technological viewpoint because it may not be up to date with themes at the forefront of research. Technology with access to up-to-date bibliographic databases is recommended instead.

Critically review the text and pay attention to the model's hallucinations.
Make sure that the information is correct, well documented and exhaustive.
Check that other authors' opinions or thoughts are quoted in a complete and accurate manner.

Generating the entire state of the art section or chapter from articles provided by the user

It complies with the University Policy principles.
It requires citing the use of GenAI.

Key points of attention

Risk of hallucinations in the case of very long documents.
Risk of wrong information.
Risk of copyright infringement.

Suggestions and actions to complete

It is recommended that you avoid providing too many documents or documents that are too long because this increases the chances of hallucinations.
Make sure that the information is correct, exhaustive and well documented.
Check that other authors' opinions or thoughts are quoted in a complete and accurate manner.

Generating the state of the art section or chapter from the user's draft

It complies with the University Policy principles.
It does not require citing the use of GenAI, unless the creative contribution made by AI prevails over the researcher's own content.

Key points of attention

Risk of wrong information.
Risk of copyright infringement.

Suggestions and actions to complete

Make sure that the information is correct, well documented and exhaustive.
Check that other authors' opinions or thoughts are quoted in a complete and accurate manner.
Carefully review the generated text.

Generating a comparison between authors or pre-existing articles to be included in an article you are writing

It complies with the University Policy principles.
It requires citing the use of GenAI.

Key points of attention

Risk that the comparisons are inaccurate.
Risk of copyright infringement.

Suggestions and actions to complete

It is recommended that you upload the texts to be compared, without assuming that GenAI knows them.

Check that the comparison between authors or articles is accurate and based on reliable sources.
Check that the comparisons are correct and the information is correctly attributed.
Make sure that other authors’ opinions or thoughts are quoted in a complete and accurate manner.
Carefully review the generated text.

Generating bibliography

Generating up-to-date bibliographies by research theme

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Risk that the generated information is incomplete because GenAI tends to find established knowledge and may overlook recent articles.
Risk that the sources found and used are inaccurate.

Suggestions and actions to complete

Using GenAI is not recommended for generation from scratch. Conversely, GenAI is appropriate for re-formatting an existing bibliography. It is recommended that you use dedicated search engines or GenAI tools that integrate up-to-date databases.

Check that the sources found are accurate, up to date, relevant and reliable.
Check that the articles and references included are pertinent to avoid including any irrelevant material.
Review the accuracy of the sources used also by means of specialist search engines (e.g. Scholar, Scopus).
Pay attention to inaccuracies, because recent articles may be left out.

Analysing data, including inferences

Using GenAI for data analysis to identify hidden patterns or insights

It complies with the University Policy principles.
Using GenAI is strongly discouraged for technological reasons. If you do not wish to follow this strong suggestion, you must cite its use.

Key points of attention

GenAI is not technically fit for this purpose.

Suggestions and actions to complete

Using GenAI is strongly discouraged because it is not technically fit for this purpose.
Rather than using GenAI, it is recommended that you train a predictive model specifically for this purpose (e.g. classical statistical or data mining approaches).

If, in spite of this suggestion, you still wish to use GenAI, you must:
Remember that the responsibility for choices and decisions based on the statistical models generated by GenAI lies with the researcher.
Apply critical thinking to the results in order to avoid making wrong decisions due to a poor understanding of data and implications.
Make sure that you have an appropriate legal basis for the processing of any personal data.
Obtain specific permission for using any data from protected sources.
Before using inferences, check that they do not introduce any bias or discrimination against individuals or groups.

Generating predictions based on historical data and current trends

It complies with the University Policy principles.
Using GenAI is strongly discouraged for technological reasons. If you do not wish to follow this strong suggestion, you must cite its use.

Key points of attention

GenAI is not technically fit for this purpose.

Suggestions and actions to complete

Remember that the responsibility for choices and decisions based on the statistical models generated by GenAI lies with the researcher.
Apply critical thinking to the results in order to avoid making wrong decisions due to a poor understanding of data and implications.
Make sure that you have an appropriate legal basis for the processing of any personal data.
Obtain specific permission for using any data from protected sources.
Before using inferences, check that they do not introduce any bias or discrimination against individuals or groups.

Making a comparative analysis to infer correlations between variables

It complies with the University Policy principles.
Using GenAI is strongly discouraged for technological reasons. If you do not wish to follow this strong suggestion, you must cite its use.

Key points of attention

GenAI is not technically fit for this purpose.

Suggestions and actions to complete

Discouraged from a technological viewpoint. GenAI is not technically fit for this purpose. Rather than using GenAI, train a predictive model specifically for this purpose using classical statistical or data mining approaches.
If, in spite of this suggestion, you still wish to use GenAI, you must:
Ensure that the comparative analysis is made correctly and based on reliable data.
Check that the inferred correlations are supported by solid evidence and results are accurately interpreted, avoiding wrong conclusions.
Make sure that you have an appropriate legal basis for the processing of any personal data.
Obtain specific permission for using any data from protected sources.
Check that the inferences do not introduce any bias or discrimination against individuals or groups.

Generating synthetic data from real data

Creating synthetic datasets retaining the statistical characteristics of the original data

It complies with the University Policy principles.
It requires citing the use of GenAI.

Key points of attention

Risk of bias in real data which is reflected in synthetic data.
Risk of privacy breach and personal data breach.
Risk of copyright infringement and confidentiality breach.

Suggestions and actions to complete

Ensure respect for privacy, including the existence of an appropriate legal basis, when data contain personal information.
Make sure that the synthetic data generated by GenAI do not reflect biases in the real data, to avoid jeopardising accuracy and any potential discriminatory effects.
Adopt a verification process for generated data statistically similar to real data.
Perform a sample check to verify that the synthetic data cannot be linked to specific individuals and comply with data protection regulations.
Check that you have specific permission to upload to the GenAI system copyrighted or confidential data, or data subject to other contractual obligations.

Enriching real datasets with generated data to test large-scale predictive models

It complies with the University Policy principles.
It requires citing the use of GenAI.

Key points of attention

Risk of bias in real data which is reflected in synthetic data.
Risk of privacy breach and personal data breach.
Risk of copyright infringement and confidentiality breach.

Suggestions and actions to complete

Ensure respect for privacy, including the existence of an appropriate legal basis, when data contain personal information.
Make sure that the synthetic data generated by GenAI do not reflect biases in the real data, avoiding jeopardising accuracy and any potential discriminatory effects.
Perform a sample check to verify that the generated data respect confidentiality and comply with data protection regulations.
Check that you have specific permission to upload to the GenAI system copyrighted or confidential data, or data subject to other contractual obligations.
Keep real and generated data separate or separable, e.g. use separate folders or archives or label real and synthetic data differently.

Generating synthetic data from generic dataset characteristics

Generating synthetic data adhering to a partially or fully defined target probability distribution (e.g. mean, variance, probability density)

It complies with the University Policy principles.
It requires citing the use of GenAI.

Key points of attention

GenAI is not technologically fit for this purpose.

Suggestions and actions to complete

Using GenAI is not recommended from a technological viewpoint. It is recommended that you use tools more fit for this purpose than GenAI, such as dedicated software for generating synthetic data based on random numbers, which allow for accurate control over the probability distribution of the generated data.
If, in spite of this suggestion, you still wish to use GenAI, you must:
Ensure respect for privacy, especially when data contain personal information.
Perform a sample check to make sure that the generated synthetic data cannot be linked to specific individuals and comply with data protection regulations.
Pay attention to the definition of target probability, considering risk factors protected under anti-discrimination regulations.

Generating synthetic data from a verbal description of the data type to be generated

It complies with the University Policy principles.
It requires citing the use of GenAI.

Key points of attention

Risk of privacy breach and personal data breach.
Risk of breach of anti-discrimination regulations.
Risk of bias.

Suggestions and actions to complete

Ensure respect for privacy, including the existence of an appropriate legal basis, when data contain personal information.
Perform a sample check to verify that the generated synthetic data cannot be linked to specific individuals and comply with data protection regulations.
Pay attention to the definition of target probability, making sure that it complies with anti-discrimination regulations and does not introduce biases linked to protected risk factors.

Manipulating data formats

Automatically converting data from a structured format into another

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Risk of errors or loss of information when merging and normalising data.
Risk of privacy breach and personal data breach.
Risk of copyright infringement and confidentiality breach.
Risk of loss of integrity and quality of the converted data.

Suggestions and actions to complete

For significant amounts of data to be converted, specific automated tools (e.g. dedicated software) are recommended instead.

Check that operations are performed correctly to maintain data integrity.

Perform a sample check to verify that conversion, merging and normalisation do not introduce errors or loss of information, ensuring that the resulting data are consistent and ready for further analysis.
Ensure respect for privacy, including the existence of an appropriate legal basis, when data contain personal information.
Check that you have specific permission to upload to the GenAI system copyrighted or confidential data, or data subject to other contractual obligations.
Ensure that data conversion does not jeopardise data integrity and quality.
Always check data integrity again in the target format.

Merging datasets with different formats into a single compatible file

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Suggestions and actions to complete

Check that operations are performed correctly to maintain data integrity.
Perform a sample check to verify that conversion, merging and normalisation do not introduce errors or loss of information and that the resulting data are consistent and ready for further analysis.
Ensure respect for privacy, including the existence of an appropriate legal basis, when data contain personal information.
Check that you have specific permission to upload to the GenAI system copyrighted or confidential data, or data subject to other contractual obligations.
Ensure that data conversion does not jeopardise integrity and quality.
Always check data integrity again in the target format.

Normalising and cleaning data from heterogeneous sources

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Suggestions and actions to complete

For significant amounts of data to be converted, specific automated tools (e.g. dedicated software) are recommended instead.
Check that operations are performed correctly to preserve data integrity.
Perform a sample check to verify that conversion, merging and normalisation do not introduce errors or loss of information, ensuring that the resulting data are consistent and ready for further analysis.
Ensure respect for privacy, including the existence of an appropriate legal basis, when data contain personal information.
Check that you have specific permission to upload to the GenAI system copyrighted or confidential data, or data subject to other contractual obligations.
Ensure that data conversion does not jeopardise integrity and quality.
Always check data integrity and consistency again in the target format.

Generating a programming code

Automatically writing functions or scripts for a specific purpose

It complies with the University Policy principles.
It does not require citing the use of GenAI, unless the contribution made by GenAI to writing functions or scripts prevails over the researcher's one.

Key points of attention

Risk of intellectual property infringement.
Risk that the generated code violates legal regulations.

Suggestions and actions to complete

Check that you have permission to use any code you input into the GenAI system.
Make sure that the generated code does not infringe intellectual property rights.
Ensure that the generated code and algorithmic solutions do not entail discrimination or violate regulations.
Always check that the generated code is syntactically and semantically correct.
Check that the generated code is consistent with the rest of the existing code, if any.

Automatically writing a programming code from a textual explanation of what the code is supposed to do

It complies with the University Policy principles.
It does not require citing the use of GenAI, unless the contribution made by GenAI to writing functions or scripts prevails over the researcher's one.

Key points of attention

Risk of intellectual property infringement.
Risk that the generated code violates legal regulations.

Suggestions and actions to complete

Explanation or documentation from a programming code

Generating detailed comments to explain the functioning of blocks of code

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Risk that the comments are inaccurate and contain errors.

Suggestions and actions to complete

Check that the comments are clear, precise and accurate.
Check that the explanations provided are correct and useful to understand how the code works.
Check that there are no errors or ambiguities in the comments that may mislead the user.

Automatically creating API documentation based on a source code

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Risk that the API documentation is inaccurate and contains errors.

Suggestions and actions to complete

Check that the documentation is clear, precise and accurate.
Check that the information provided is correct and useful to understand how the code works.
Check that there are no errors or ambiguities in the documentation that may mislead the user.

Step-by-step explanation of the logics implemented in an algorithm

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Risk that the explanations are inaccurate and contain errors.

Suggestions and actions to complete

Check that the information is clear, precise and accurate.
Make sure that the explanation reflects the algorithmic solutions correctly.
Avoid simplification that may distort how the algorithms actually work.

Generating diagrams (e.g. UML) illustrating how a code works

It complies with the University Policy principles.
It requires citing the use of GenAI, if the entire image is generated.

Key points of attention

Risk that the diagrams are inaccurate.

Suggestions and actions to complete

It is recommended that you generate a design language code (e.g. PlantUML) rather than images directly.
Check that the generated diagrams are accurate and illustrate how the code works.

Instructions on using a code or software

Explaining how a block of code or software module works

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Risk that the explanation is inaccurate.

Suggestions and actions to complete

Check that the explanation is clear, accurate and comprehensible.
Check the code's licence to make sure that the use complies with contractual terms and conditions.

Querying GenAI as an alternative to consulting a software or application manual or documentation

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Risk of inaccuracy.
Risk of breach of software use terms and conditions.

Suggestions and actions to complete

Check that the explanation is clear, accurate and comprehensible.
Check the code's licence to make sure that the uploading of the documentation complies with use terms and conditions.

Image generation

Automatically creating logos based on specific requests or themes

It complies with the University Policy principles.
It requires citing the use of GenAI.

Key points of attention

Risk of intellectual property infringement.
Risk of inaccuracies.

Suggestions and actions to complete

Carefully check that automatically creating logos does not inadvertently infringe copyright or registered trademarks.
Check the precision of details in the generated images, such as lettering and other elements.

Generating illustrative images to accompany papers, articles or presentations

It complies with the University Policy principles.
It requires citing the use of GenAI.

Key points of attention

Risk that visual graphics are inaccurate or not true to the data or research results.
Risk of generating deep fakes.
Risk of privacy breach and personal data breach.

Risk of copyright infringement.

Suggestions and actions to complete

Check that visual graphics represent data or research results in an accurate, clear and true manner, avoiding distortions that jeopardise understanding or interpretation.
Make sure that charts and visualisations are correct and well documented.
Only use unrestricted or duly authorised materials, avoiding intellectual property infringements or privacy breaches.
Make sure that you have an appropriate legal basis for the processing of any personal data (e.g. images of real people).
Check that you do not violate personal data protection and image rights, as synthetic images may be likely depictions of real people.
Make sure that the generated images do not replicate recognisable elements of protected works to avoid copyright infringements.
Check that the generated images are relevant, correct and comply with privacy and copyright laws.
Avoid generating and disseminating content that depicts real people doing things that did not actually happen, as this may classify as a deep fake and manipulate public opinion.
When "evidently creative" content is synthetic, clearly state this, in accordance with Article 50 of the AI Act.
Check the precision of details in the generated images (e.g. lettering).

Linguistic assistant

Automatically correcting grammatical and stylistic errors in academic documents

It complies with the University Policy principles.
t does not require citing the use of GenAI, unless such use is the aim of research itself.

Key points of attention

Risk of language bias.
Risk of copyright infringement.

Suggestions and actions to complete

Check that machine translation does not introduce any language bias, including gender bias, to ensure that the content is gender-neutral and inclusive.

Make sure that you have the necessary authorisations to translate the material.

Translating technical texts or scientific articles into several languages

It complies with the University Policy principles.
It does not require citing the use of GenAI, unless such use is the aim of research itself.

Key points of attention

Risk of language bias.
Risk of copyright infringement.

Suggestions and actions to complete

Check that machine translation does not introduce any language bias, including gender bias, to ensure that the content is gender-neutral and inclusive.
Make sure that you have the necessary authorisations to translate the material.

Suggestions to improve the clarity and coherence of a text

It complies with the University Policy principles.
It does not require citing the use of GenAI, unless such use is the aim of research itself.

Key points of attention

Risk of language bias.
Risk of copyright infringement.

Suggestions and actions to complete

Knowledge valorisation

Generating presentations

It complies with the University Policy principles.
It requires citing the use of GenAI.

Key points of attention

Risk that the content is inaccurate.
Risk of copyright infringement.

Suggestions and actions to complete

Check that the generated presentation is accurate.
Check that the content is correct, relevant and well structured.
Check that the generated presentation contains no errors or inaccuracies that may jeopardise the quality and clarity of the presentation itself.
Make sure that you have the necessary authorisations to upload the text to the GenAI system.

Summarising scientific texts for informative purposes

It complies with the University Policy principles.
It does not require citing the use of GenAI.

Key points of attention

Risk that the summary is inaccurate.
Risk of copyright infringement.

Suggestions and actions to complete

Check that the summary is accurate and comprehensible.
Check that the scientific content is simplified without distorting key concepts.
Critically review the generated output to maintain the integrity of information while adapting it to the level of understanding of the audience.
Make sure that you have the necessary authorisations to upload the text to the GenAI system.

You can share observations, suggestions for additional use cases, or modifications to the reported use cases that you find most relevant to the University Policy and its application. Share your feedback.