Authorised AI tools

Policy overview

Agile Collective have several options available:

  • a business account for Claude (available to ALL members on request)
  • chat.agile.coop (available to all employees)

If another (paid for) service is required, then this will be considered on an adhoc basis by the AI Circle to confirm the need and strategy around additional tools. If you use another service, you are responsible for inserting the charge into the AI spreadsheet

You are also responsible for observing the "rejected tools" list on this page before using something new.

NOTE: There are many areas of AI in the world, and this document is trying to stay broad so we can review all relevant AI tooling. This has meant that the term "AI tool" has been used throughout, but can refer to a variety of things such as: AI Providers, specific AI models, general AI features, etc. Where possible, we will try to be specific, but in most cases this wont be possible on this page.

FAQs

How do we determine whether to use an "AI tool"?

To determine whether an "AI tool" is fit for use, the AI Circle should be sent all useful information in a timely manner, so they can review it.

The key areas used in this review are:

  1. Hardware location : Are there country specific risks to consider, in most cases, UK based servers are preferable
  2. Green credentials : Do they use renewable energy for example, and do they share their environmental data publicly.
  3. Zero data retention policy: The ability to configure zero data retention by the provider as required.
  4. Ethical standing : Such as military contracts and general newsworthy ethical considerations.

If you wish for an "AI tool" to be reviewed, sharing information and thoughts on these key areas would speed up the process.

What if tool x is not in the list?

You must submit a request for review to the AI Circle.

When doing so, it would really help if you could pay particular attention to the 3 key areas outlined above, supplying any related docs if you have them to hand.

Can I use tool y even though it has been rejected?

When a tool has been rejected, the AI Circle have deemed it either unsafe, or against Agile Collectives ethics.

Therefore, if a tool has been rejected, then you must not use/install it on an Agile Collective device (phone/laptop), or utilise it to work on any Agile Collective work (both internal and external).

How long can it take for a tool to be approved?

This depends on when it was submitted, the availability of the team, and the complexity of the tests. At the moment, it could take hours/days/week to review, so please submit in a timely manner. Once a few reviews have been undertaken, we may be able to estimate more accurately.

Tools: Generative AI (GAI)

Note: When an "AI tool" is added to the list it should include a little background/detail on what it is, and why the decision was made, and any context on what the decision was for.

These are the main "AI tools" members at Agile collective are using, and are more knowledgeable about

Approved Tools

The below Tools are ones that have been reviewed and considered okay to use, but are not necessarily the core tools for Agile Collective.

Rejected Tools

Name Reason
Grok (developed by xAI) Rejected due to ethical alignment

Last updated: