Government Analysis Function > Artificial Intelligence (AI) Hub > Artificial Intelligence (AI) glossary and prompt library

Artificial Intelligence (AI) glossary and prompt library

Artificial Intelligence (AI) is often discussed using technical language, which can make it harder to engage with. This glossary provides clear, plain-English explanations of common AI terms used across this hub and in the wider Civil Service Learning.

You can use this glossary to:

  • look up unfamiliar terms as you work through the guidance or learning
  • build confidence when talking about AI
  • as a reference resource whenever you need a reminder

No technical background is required to use the glossary. Please note: this list of terms is not exhaustive, and we have added some external links to other A to Z glossaries on this page.

Term descriptions

AI agents

Small computer programmes that can perform tasks on your behalf.

AI model

A ‘map’ of possible connections between data. For example, in a Large Language Model (LLM), the connections will be between words or units of language. When the model receives input from a prompt, it explores those potential connections to generate the most relevant and plausible output.

Anthropomorphising

The act of attributing human traits, emotions, or intentions to non-human entities. For example, saying an AI “thinks” or “understands”.

Artificial Intelligence (AI)

Technology that makes computers more useful and capable of performing complex tasks on their own.

Alternative definition

The AI Playbook for the UK Government uses the definition of AI adopted by Organisation for Economic Co-operation and Development (OECD) countries:

“An AI system is a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment.”

Bias

In the context of AI, bias refers to systematic errors in an algorithm that lead to unfair or inaccurate outcomes. This often occurs when the data used to train the AI reflects existing human prejudices or is not representative of the population it will affect.

Big Data

Extremely large and complex datasets that cannot be easily managed or analysed with traditional tools.

Chatbots

AI personal assistants that can help with the summarising, searching, and writing of information.

Conversational AI

Technologies, such as chatbots or virtual agents, that users can talk to. They use large volumes of data, machine learning and natural language processing to help imitate human interactions, recognising speech and text inputs, and translating their meanings across various languages

Data

Data is a collection of facts, numbers, words, observations or other useful information. Data can exist in various forms, such as numbers, text, images, or measurement. Not all chatbots are equipped with AI, but modern chatbots increasingly use conversational AI techniques such as natural language processing (NLP) to understand user questions and automate responses to them.

Dataset

A structured collection of individual pieces of data.  Typically organised in a table or other structured format, datasets are used to train and test AI models.

Hallucinations

When AI makes up information, also known as ‘fabrications’.

Input

Data given to the AI to create an output

Large Language Models (LLMs)

AI systems that use machine learning to help them process language so that they can mimic the way humans communicate.

Machine Learning

A field of computer science under the umbrella of AI where people teach a computer system how to do something by training it to identify patterns and make predictions based on what it has learned.

Natural Language Processing (NLP)

NLP enables computers and digital devices to recognise, understand and generate text and speech by combining computational linguistics, the rule-based modeling of human language together with statistical modeling, machine learning and deep learning.

Output

The response generated by AI.

Prompt

The question you ask to AI.

More A to Z lists of AI terms

Here are some other useful glossaries and lists of AI terms:

Prompt library

This prompt library brings together a collection of tested instructions, templates, and text recipes designed to help you get precise and repeatable results from generative AI tools.

“You are an expert awards writer specialising in the Analysis in Government (AiG) Awards. Before undertaking this task, review the following official guidance: 

Use these pages as the definitive source for: 

  • award categories and criteria 
  • what judges are looking for 
  • AiG Awards writing best practice 
  • the style, structure and evidence expected in successful nominations 

I will provide information about an individual, team, project, piece of analytical work, programme or achievement. 

The information may be incomplete and may come from multiple sources such as: 

  • project documentation 
  • evaluation reports 
  • presentations 
  • business cases 
  • performance metrics 
  • testimonials 
  • lessons learned 
  • meeting notes 
  • emails 
  • impact assessments 
  • bullet-point notes 

Your task is to: 

Step 1: Understand the work

Analyse all information provided and identify: 

  • the challenge, problem or opportunity addressed 
  • the analytical contribution made 
  • main actions taken 
  • stakeholders involved 
  • users and beneficiaries 
  • evidence of innovation 
  • evidence of collaboration 
  • evidence of leadership 
  • evidence of inclusion 
  • evidence of capability building 
  • outcomes and impact 
  • quantifiable benefits and measures of success 

Step 2: Category assessment

Compare the evidence against every AiG Award category. 

For each category: 

  • provide a suitability score out of 10 
  • explain which criteria are met 
  • identify any criteria that appear weaker or unsupported 

Recommend the single strongest AiG Award category and explain your reasoning. 

Step 3: Identify gaps

Identify: 

  • missing evidence 
  • missing metrics 
  • areas requiring clarification 
  • potential weaknesses a judging panel may identify 

Provide a list of suggested information that would strengthen the nomination. 

STEP 4: Draft the nomination

Using only the evidence provided, draft a complete AiG Award nomination. 

Do not invent information, achievements, benefits or impacts. 

If information is missing, make a note of where further evidence would strengthen the submission. 

Produce the following: 

Part A: Nomination summary

Maximum 150 words. 

The summary should: 

  • explain what was achieved 
  • explain why it matters 
  • highlight the most significant impact 
  • capture the judges’ attention 
Part B: Full nomination description

Maximum 750 words. 

Structure under the following headings: 

  • challenge or opportunity 
  • action taken 
  • evidence of excellence 
  • outcomes and impact 
  • alignment to the Award Criteria 

The writing should: 

Step 5: Judge’s review

Act as an AiG Awards judge and provide: 

  • overall assessment 
  • main strengths 
  • potential concerns 
  • expected competitiveness 
  • five specific improvements that would increase the nomination’s chances of success 

Step 6: Quality assurance check

Confirm: 

  • the nomination is within AiG word limits 
  • all claims appear supported by the evidence provided 
  • no obvious exaggerations or unsupported assertions have been introduced 
  • any areas that should be fact-checked before submission 

Source material: 

[Paste supporting information into your chosen AI tool.]”

“You are an expert awards consultant specialising in the Analysis in Government (AiG) Awards.

Before analysing the nomination, carefully review the official AiG Awards guidance:

Use the information on these pages as the definitive source for:

  • award category descriptions
  • award criteria
  • what judges are looking for
  • best practice for writing successful nominations
  • the style, structure and evidence expected in AiG Awards submissions

I will provide an award nomination originally written for a different awards scheme.

Your task is to:

  1. Analyse the nomination.
  2. Compare it against all AiG Award categories and criteria.
  3. Score suitability against each category.
  4. Recommend the best-fit AiG Award category.
  5. Explain your reasoning.
  6. Identify any gaps, weaknesses or missing evidence.
  7. Suggest specific improvements that would strengthen the nomination against the chosen category.
  8. Redraft the nomination to maximise success in the recommended category.

When carrying out the analysis:

  • focus on evidence, outcomes and impact rather than activity
  • prioritise measurable results where available
  • highlight analytical excellence, innovation, collaboration, leadership, capability building, inclusion and user impact where relevant
  • remove unnecessary background information and duplication
  • make explicit links to the chosen AiG Award criteria
  • follow the advice on the AiG “Writing a Winning Nomination” page throughout

Do not invent, exaggerate or assume achievements that are not evidenced in the original nomination. If important information appears to be missing, identify the gap and indicate where additional evidence would strengthen the nomination.

Output

Part A: Category mapping
  • summary of nominated work
  • suitability score for each AiG Award category
  • recommended category
  • reasons for recommendation
  • strengths against the criteria
  • gaps and opportunities for improvement
Part B: Nomination summary

Maximum 150 words.

The summary should:

  • clearly explain what was achieved
  • explain why it matters
  • highlight the most significant impact
  • encourage judges to want to read more
Part C: Full nomination description

Maximum 750 words.

Structure the nomination around:

  • challenge or opportunity
  • action taken
  • evidence of excellence
  • outcomes and impact
  • alignment to the chosen AiG Award criteria 

The writing should be persuasive, evidence-based, concise and suitable for assessment by senior analytical leaders and civil servants serving as judges.

Part D: Quality assurance check

Finally, provide:

  • confirmation that the nomination stays within the AiG word limits
  • any factual claims that require verification
  • any areas where additional evidence, metrics or testimonials would strengthen the submission
  • a final assessment of how competitive the nomination appears against the published criteria

Original nomination:

[Paste the original awards nomination previously used for another awards scheme.]

Return to AI Hub homepage. 

Related