AI-generated requests are increasing, and I don't think it's a bad thing

May 21
2 min read

Organisations should not dismiss or treat requests differently because AI may have been involved.

Anyone who knows me knows that I am super passionate about information rights and the role they play in protecting people. At their core, both FOIs and SARs are about empowerment. They are intended to give people the ability to understand decisions, challenge organisations, access their own information, and hold public bodies to account - often where there is a pronounced power imbalance.

So, when people use AI tools to help them write requests, my instinct is not to see that as a threat but something that can help people who may previously have struggled to articulate what they wanted to ask for. It can also improve confidence, accessibility, and engagement with rights that have often felt intimidating or difficult to navigate.

That said, we also need to be honest about how this impacts our work. We are starting to see more FOI requests and SARs that appear to have been drafted using generative AI, and some of these requests are much longer, broader, and more complex than traditional requests. They can contain multiple overlapping questions, very wide or vague wording, or template-style language that is not always tailored to the organisation receiving it.

More time may be needed to actually interpret the request, clarify scope, search systems, review material, and coordinate responses across different services. In SAR handling particularly, wider requests can increase the volume of personal data that needs to be reviewed and redacted before disclosure. The challenge is managing this increasing complexity within stretched teams.

The ICO’s new guidance on AI-generated FOI requests is helpful because it reinforces a balanced and proportionate approach. The guidance makes clear that a request does not become invalid simply because AI was used to draft it. The same legal principles still apply. Organisations should not dismiss or treat requests differently because AI may have been involved.

At the same time, the ICO recognises that organisations can still use the existing mechanisms available to them. That includes asking for clarification where requests are unclear or excessively broad, considering proportionality and cost limits where relevant, and documenting decisions properly. In reality, those same principles are just as relevant for SAR handling as they are for FOI.

For me, the answer is not to push back against AI-assisted requests. It is to make sure my customers respond in a calm, fair, and practical way. We need processes that protect people’s rights while also helping teams manage workload sustainably. Clear communication, confident decision-making, and consistent approaches across FOI and SAR handling will become increasingly important as AI use continues to grow.

Ultimately, if AI helps more people exercise their information rights, i'm all in. My focus now is just making sure we are equipped to respond well.