1. POLICY STATEMENT
This policy concerns the use of Artificial Intelligence (AI) including generative AI, such as Chat GPT, by staff
in our organisation. AI tools are becoming widely available. They can answer questions quickly in a variety
of styles and formats and are increasingly being used for research, drafting, interaction with customers and
recruitment and selection. This policy aims to make clear the principles our organisation will adopt when
using, or making decisions about using, AI. All staff must adhere to these principles and seek to apply them
in practice.
2. IS THIS POLICY PART OF MY CONTRACT OF EMPLOYMENT?
No, and we can change this policy at any time, but if any changes are made, we’ll always make you
aware of them. We may also vary things like time limits in this policy, if we feel we need to.
3. WHO’S COVERED BY THIS POLICY?
This policy applies to all employees, directors and other officers, workers and agency workers. We also
require in any contracts with self-employed consultants or contractors that they comply with this policy, and
we’ll make sure they’re given access to a copy. All these people are referred to as ‘Staff’ in this policy.
4. WHO IS RESPONSIBLE FOR THIS POLICY?
While we ask all managers to take responsibility for making sure this policy is complied with, its successful
operation also depends on you. Please take the time to read and understand it, and to go back to your
manager with any questions you may have. References to Directors in this policy mean the most senior
people within our organisation.
5. WHAT DO THE TERMS MEAN?
5.1 Generative AI
“Generative Artificial Intelligence” is a type of AI which can be used to create new material or output, such
as text, code, images, music.
5.2 Large Language Models (LLMs)
LLMs are a type of generative AI that can produce human-like text in response to a prompt. They use deep
learning techniques and large volumes of data.
5.3 GPT
This means “generative pre-trained transformer” and is a type of LLM that uses deep learning to produce
texts in natural language based on information requested.
5.4 Prompts
Prompts are the inputs or queries that a user provides to the generative AI application to generate an
output.
6. WHAT SHOULD I KNOW ABOUT THE USE OF AI?
6.1 Accuracy
The accuracy of any output generated by AI will depend on the data which the AI was “trained” on. As a
result, it can be inaccurate and can magnify any inaccuracies. This is known as “hallucinating” – the ability
of generative AI to create plausible, but entirely incorrect, outputs.
6.2 Bias
AI will perpetuate any built-in biases contained in the data it was trained on and of those who set the
system up. Furthermore, AI will have a tendency to reinforce these inbuilt biases, depending on how the
algorithm works.
6.3 Breach of Intellectual Property (IP) Rights
AI is likely to reproduce material that’s subject to intellectual property rights, including patents and
copyrights, and these may not be obvious.
6.4 Confidentiality & Data Protection Issues
Where users enter data, generative AI may add this data to its database and make further use of it. If the
data entered is personal data, then passing on the data in this way is likely to be a breach of GDPR unless
there has been full transparency when the data was collected.
6.5 Data Sharing
All internal data, as well as data shared with Transform from our clients, must not be uploaded to AI.
Outputs generated by generative AI may themselves contain personal data that has not been processed in
accordance with GDPR.
7. OUR PRINCIPLES ON USING AI
Because the possibilities of generative and other forms of AI are developing so quickly and are likely to
have impacts that we cannot predict, we’ll use a principles-based approach to their use within our
organisation.
All staff are responsible for applying these principles at all times when using AI. Because we’re all learning
together, and because we know you’ll encounter unfamiliar situations, it’s important that you discuss any
questions or issued with your manager as soon as possible.
Our principles on using AI are:
a) We are transparent about our use of AI;
b) We are responsible for all outputs generated by AI;
c) We are responsible for the accuracy of all outputs generated by AI;
d) We understand that AI can be biased and can magnify inbuilt biases and we will constantly work to identify
and eliminate these biases;
e) We will not operate AI in a way that breaches other people’s Intellectual Property (IP) rights;
f) We will operate AI strictly in accordance with data protection law and in particular individuals’ rights under
data protection law;
g) Where our customers, job applicants, staff and others are affected by decisions made by AI, they will have
the right of a review by a human.
h) Internal data and data shared with Transform from clients will not be uploaded to AI.
8. HOW CAN I USE AI IN THE WORKPLACE?
We’ll let you know which AI apps you’re allowed to use for work, and any we don’t allow you to use.
You must always select the “opt-out” option when you use AI, to prevent data and inputs being used by the
AI to train itself.
You must not input offensive or discriminatory content as a prompt.
You may use AI for the following purposes:
a) Drafting internal materials including presentations and papers
b) Research
c) Producing marketing materials
d) Generating ideas
e) Producing instructions
f) Writing code
You must always disclose when any work you produce has been created by AI.
9. CAN I USE AI TO RESEARCH TOPICS?
We believe AI will become an increasingly useful and powerful tool for background research and has a
significant role to play in saving time.
However, like all tools it should be used with care.
a) Outputs generated by AI should be carefully checked for accuracy;
b) Generated outputs must be credited as produced by an AI tool and not passed off as your own work.
10. CAN I USE AI TO PRODUCE MATERIALS?
We believe AI could be used to draft a wide range of internal and customer-facing materials, including
policies, advertising materials, instructions, procedures, etc.
While AI-generated output may be used as a starting point, staff are responsible for ensuring it complies with
our principles and in particular those on accuracy, bias, IP rights and data protection.
Staff are responsible for the outputs they produce.
11. HOW WE USE AI IN RECRUITMENT & SELECTION
Where we use AI in making recruitment and selection decisions, we’ll make that clear to applicants in our
Privacy Notice and give applicants the right for any decision to be reviewed by a human.
12. WHAT OTHER POLICIES ARE RELEVANT?
Our policy on the use of AI isn’t confined to this policy alone. Other policies that are likely to be relevant are:
a) Our policy on Data Protection
b) Our policy on using IT Equipment, Information Systems and Data
c) Our policy on Equality, Diversity and Inclusion
d) Employee Code of Conduct
13. WHAT IF I BREACH THIS POLICY?
If you breach this policy, you may face disciplinary action and possible dismissal.
14. MONITORING AND REVIEW OF THE POLICY
We will continue to review the effectiveness of this policy to ensure it is achieving its stated objectives
