Artificial intelligence (AI) guidance and information sharing policy

1. Purpose and scope

We acknowledge the increasing use of artificial intelligence (AI) tools in content production, including research and writing processes. We also recognise that AI can offer opportunities to improve efficiency, enhance collaboration, and support small or remote teams. However, its use must be balanced with legal compliance, maintaining BISA’s academic and professional standards, and ethical considerations.

This guidance sets out how AI, including Generative AI (GenAI), may be used in relation to BISA activities. The purpose is to:

• Protect BISA from legal, financial, and reputational risk 
• Promote responsible, ethical, and transparent use of AI 
• Support staff, trustees, and contributors in using AI effectively.

These guidelines apply to all individuals producing content for or on behalf of BISA, including:

• Staff and trustees 
• Working group conveners 
• Contributors of content to BISA (e.g. websites, blogs, reports, papers, events, videos and images).

This is guidance does not apply to submissions and related materials to either BISA journal – Review of International Studies and European Journal of International Security. Please see their respective websites for their individual policies on AI.

2. What counts as AI use?

For the purposes of this guidance, AI use includes (but is not limited to):

• Generating or editing written content (e.g. blogs, event descriptions, minutes, news stories) 
• Creating, modifying or enhancing images, audio, or video
• Analysing data
• Drafting social media posts.

“Content” in this context includes:

• Text 
• Images and graphics 
• Audio and video 
• Data outputs.

3. Key risks

Legal/financial - Use of AI may expose BISA to:

• Copyright infringement
• Inaccurate or misleading information (“hallucinations”) 
• GDPR and personal data breaches 
• Uploading of confidential or sensitive information into AI systems.

These risks may carry legal and financial consequences for BISA and are a primary concern.

Reputational - BISA is a scholarly association and wants to publish content that is authentic and engaging. AI-generated content may appear generic, inaccurate, or inauthentic, and overuse could undermine BISA’s credibility. In addition, poor quality outputs may reduce accessibility - especially for non-native English speakers.

Skills development - Overreliance on AI may erode critical skills such as critical thinking, writing, communication and creativity. Read ‘Your brain on ChatGPT’ for more information.

Environmental and other ethical concerns - AI systems have significant environmental costs (such as energy and water usage). There are also other ethical concerns including, but not limited to, AI models being trained using stolen data, and low-paid workers in the Global South being used to build AI. In the context of BISA’s sustainability goals and position as a registered charity, these indirect costs should be considered when deciding whether AI use is necessary.

4. When AI is, or is not, appropriate

It may be appropriate to use AI where it supports efficiency and quality without replacing human judgment. However, decisions should be made on a case-by-case basis, taking account of the considerations in section 3 of this guidance. AI should only ever be used as a supporting tool, not the primary author.

If you are unsure whether your use case is appropriate, please discuss this with BISA Communications and Marketing Director, Chrissie Duxson: Chrissie.Duxson@bisa.ac.uk in the first instance.

Examples of potentially appropriate use*:

• Minimal and non-generative uses of AI tools. For instance, basic spelling or language checking, incorporation of single words, or minor formatting tasks such as converting section headers to bold type
• Editing and refining existing BISA materials that are already in the public domain
• Refining draft versions of content (with substantial human editing) 
• Improving accessibility - e.g. first drafts of alt text for images or closed captions for videos which are then checked for accuracy and appropriateness by a human before publishing
• Enhancing audio/video/image quality (where you/BISA own the copyright of source materials, or have the permission of the copyright owner)
• Supporting idea generation or brainstorming
• Analysing anonymised data (in compliance with GDPR)

*For all these use cases please consider whether you are granting the owners of the AI you are using a licence to use/distribute your work. This is particularly important when using free AI tools, but may also apply to paid tools.

AI must not be used:

• To generate content presented as entirely original human work without disclosure
• To process personal, confidential, or sensitive data
• To produce content that has not been fact-checked and verified 
• To generate material that may breach copyright
• As a substitute for expertise in scholarly or professional outputs.

If you choose to use AI, it is essential that all outputs are checked for: 

• Accuracy 
• Bias 
• Appropriateness.

5. Transparency and disclosure

BISA will adopt a standardised disclosure approach. All contributors will be asked to confirm whether AI was used in producing content using a form.

If yes, contributors must provide the following information:

• Why AI was used 
• What it was used for 
• Which tools/platforms were used 
• What prompts or inputs were provided (where possible).

This approach ensures transparency, consistency, and that no individual is singled out.

Appendix 1 – Practical checklist for contributors

This is a quick, common-sense guide to help you decide whether and how to use AI when creating content and processing data for BISA.

Before you use AI

Ask yourself:

• Do I actually need AI for this task, or would it be just as quick to do it myself? 
• Am I about to input anything confidential, personal, or sensitive?
  → If yes, do not use AI.
• Is this something that requires my own expertise or judgement?
  → If yes, AI should only support, not replace your input.

If you do use AI

Make sure you:

• Use it for support, not as a shortcut to produce finished work 
• Avoid copying and pasting outputs directly without editing 
• Keep prompts and inputs generic (no sensitive data).

After using AI

Before submitting anything to BISA:

• Have I checked all facts and references? 
• Have I rewritten and edited the content to ensure the intended objectives come across?
• Does this sound like something BISA would publish? 
• Is the tone clear and accessible (especially for non-native English readers)?
• Am I confident this doesn’t breach copyright? 

Disclosure

If you’ve used AI, be ready to briefly explain:

• What you used it for (e.g. first draft, editing, ideas)
• Which tool(s) you used 
• Why you used it 

This isn’t about catching people out — it’s about being consistent and transparent.

Appendix 2 - Common AI tools (by use case)

This is not an approved list - just a guide to what people commonly use, so expectations are clear.

Writing and editing

• ChatGPT 
• Microsoft Copilot
• Google Gemini 
• Claude

Social media and communications

• Canva (AI-assisted captions and design) 
• Hootsuite/Buffer (AI scheduling suggestions, first drafts of posts, alt text, testing tone and messaging)
• TikTok (video editing, AI captions)

Images, video, audio and design

• Canva (poster design within brand guidelines, background remover, creating clips) 
• WondershareFilmora (audio cleanup, transcription) 
• Adobe Premiere Pro (AI enhancements)
• CapCut (auto captions, editing tools)

Data and analysis

• Excel (Copilot features) 
• Google Sheets 
• ChatGPT

Collaboration

• ChatGPT 
• Miro (AI brainstorming tools)
• Notion AI

This policy was drafted and implemented in July 2026. It is due for review in July 2027, or earlier if necessary.