The submission and publication of text, images, figures, or graphs created by artificial intelligence, machine learning, language model, generative artificial intelligence, or similar algorithmic technology (hereafter referred to as an “AI tool”) is discouraged at all SSA journals, unless they are part of a study’s formal research design or methods. For the purposes of these guidelines, the term “AI tool” does not include spelling, grammar and referencing tools used in word-processing software.
If AI tools are used in the preparation of a manuscript submitted to an SSA journal, authors must follow SSA’s guidelines regarding Authorship, Disclosure, and Responsible Research.
Authorship and AI:
No AI tool will be accepted as an author on a manuscript submitted to an SSA journal. AI tools cannot meet SSA’s journal requirements for authorship, such as taking responsibility for the submitted work, reporting on conflicts of interests or managing copyright and license agreements. As with all manuscripts published by SSA, authors are fully responsible for the integrity of this content and are liable for any breach of SSA publication ethics guidelines, even if an AI tool is used to create or to assist with the production of manuscript content.
Disclosure and AI:
Authors submitting a manuscript to an SSA journal must report any use of AI tools, including those used in the collection and analysis of data or in the creation of text, images, figures, or graphs, but not including spelling, grammar and referencing tools used in word-processing software.
Such disclosures should appear in the manuscript’s Data and Resources section. Disclosures should include a description of what content was created using an AI tool, how it was created, and the name (along with version and extension numbers, if applicable) of the tool and its author or manufacturer.
In situations where AI tools have made significant changes to a manuscript’s writing style and grammar, the submission cover letter should demonstrate the authors’ understanding of any changes and describe what steps the authors have taken to validate such changes.
Responsible Research and AI:
Scientific research publications traditionally include documentation on the data, techniques, and tools used in the research, which allows for the validation and replication of any scientific results. This can be complicated when AI tools are used because computation, data processing, and/or data analysis conducted by AI tools may not be easily accessible to human researchers.
Responsible research using AI tools should be as transparent and accessible as possible. In submitted manuscripts, researchers should disclose and describe processes of data collection, data preprocessing, AI model construction, and model training used in the study. Methods that address potential bias, risk and harm resulting from AI tool use, if applicable, should also be considered in these studies.
Data analyzed with AI tools also should be transparent and accessible. Authors should disclose which datasets were used and make them available to others. Authors should also alert others to any restrictions on data access or known biases in the data.