Purpose
Info Latinos may use artificial intelligence (AI) to help with routine tasks such writing grant applications, as brainstorming, drafting, summarising public information and improving accessibility. AI must support—not replace—human judgement, relationships and responsibility.
This policy applies whenever trustees, staff, consultants or volunteers use an AI tool for Info Latinos work, including tools built into other software.

Our golden rule
Protect people and confidential information. Never enter identifiable beneficiary, volunteer, donor, safeguarding, immigration, health, family-law or other case information into a general AI tool.

For this policy, a “general AI tool” means an AI service that has not been specifically approved by Info Latinos for processing organisational information. Being built into software we already use does not automatically make an AI tool approved.

​

What this means in practice

• Do not enter names, contact details, photographs, recordings, identification documents, case notes or exact personal circumstances.

• Removing a name is not enough if the person could still be recognised from the remaining details.

• Use fictional, aggregated or genuinely anonymised examples wherever possible.

• Do not copy confidential partner, contract, funding or commercial information into a general AI tool.

• If you are unsure whether information is safe to use, stop and ask the Founder.

 

Using AI responsibly

1. Use AI for appropriate tasks
AI can be useful for first drafts, ideas, agendas, non-sensitive communications, formatting, and summaries of public or properly anonymised information.

2. Check everything
AI can invent facts, quotations, statistics, links and legal requirements. Treat every output as a draft. Check important information against reliable sources before using or sharing it.
Legal, immigration, safeguarding, medical or financial information must be checked by someone appropriately qualified.  Raw or unverified AI output must never be passed to beneficiaries as professional advice.  Any information should be passed to beneficiaries with web references including from government sites.

3. Look for bias
AI can reproduce stereotypes or work less well for some groups and languages. Check whether an output excludes, stereotypes or disadvantages people because of nationality, ethnicity, language, immigration status, age, sex, disability, religion, sexual orientation or other characteristics.
Important Spanish or other translations must be checked by a competent bilingual person, particularly where they concern rights, services or personal circumstances.
AI translation alone should not be relied upon where an error could materially affect a person’s rights, safety or access to essential services. 

4. Keep a human responsible
The person using AI remains responsible for the final work. A named person must review and approve anything that will be published, sent to beneficiaries or used to inform an important decision.
AI must never make the final decision about safeguarding, casework, access to support, volunteer selection, complaints, disciplinary matters or any other decision that significantly affects a person.

5. Be open where it matters
Routine help with spelling, formatting, brainstorming or an early draft does not usually need to be declared. Be open about significant AI use where it could affect trust, interpretation or accountability—for example, in research, service delivery or substantive published work.

 

Important
Do not use AI to imitate a real person, create a false testimonial, or generate a realistic image or voice of an identifiable person without explicit permission and safeguards.

​

Do not assume that AI-generated text, images, audio, video or other content is automatically free from copyright or intellectual-property restrictions. Check permissions where necessary before publishing or reusing it. 

 

Higher-risk uses
Staff and volunteers must obtain the Founder’s approval before:

• using any personal or sensitive data with an AI system;

• connecting AI to an Info Latinos email account, database, files or other system;

• using AI directly in services or to communicate automatically with beneficiaries;

• using AI meeting transcription, recording or note-taking where personal or confidential information may be discussed;

• allowing an AI system to take actions on behalf of Info Latinos.

 

Approval should confirm that the use is necessary, lawful and secure; that the supplier’s privacy and data-processing terms are suitable; and that clear human oversight is in place. A formal data-protection assessment may be required for higher-risk uses.

 

If something goes wrong
If you accidentally enter sensitive information, receive a harmful or discriminatory output, or suspect information has been exposed:

• stop using or sharing the output;

• tell the Founder immediately; and

• follow Info Latinos’ data-breach or safeguarding procedure where relevant.

Suspected AI-related data incidents should be recorded and assessed under Info Latinos´ existing data-breach procedure. 
Do not hide mistakes. Reporting them quickly helps protect people and improve practice.

​

Quick check before using AI

• Does this contain personal, confidential, safeguarding or case information?

• Could someone be identified even if I remove their name?

• Who will check the result—and do they have the right knowledge and language skills?

• Could the output be inaccurate, biased or harmful?

• Could there be copyright or confidentiality concerns?

• Would I be comfortable explaining this use to the person concerned or to our trustees?

• Is AI really needed, or is there a safer and simpler way?

​​

​Responsibility and review
Trustees approve this policy. The Founder approves tools and higher-risk uses, supports staff and volunteers, and responds to incidents. Everyone using AI for Info Latinos must follow the policy, check their work and raise concerns promptly.
Staff and volunteers who use AI for Info Latinos work should receive appropriate guidance on this policy, confidentiality, data protection and safe AI use.
This policy should be read alongside Info Latinos’ data protection, confidentiality and safeguarding policies. It will be reviewed annually, or sooner if our use of AI changes significantly.

Further guidance: Information Commissioner’s Office guidance on AI and data protection; Charity Commission guidance on charities and AI; and National Cyber Security Centre guidance on secure AI use.

​​​