Glossary 

AI: Artificial Intelligence, or computer systems that are able to replicate human-like intelligence tasks such as speech recognition, inferring information, decision-making, and translation. 

Algorithm: A series of logical steps or instructions. When discussed in computing, it generally means the logical steps a piece of software takes to achieve a desired outcome. 

API: Application Programming Interface, a type of software that allows communication between two or more otherwise disconnected computer programs. 

Artificial general intelligence: a hypothetical AI that is able to autonomously think like a human or even exceed the intellectual capacity of humans. Usually abbreviated to AGI to avoid confusion with AI or generative AI. 

Beta (computing): an early version that is released for testing to a limited audience. 

Bing: Microsoft's inbuilt search engine in the Edge internet browser. In February 2023, a version of Bing incorporating ChatGPT was launched in beta mode. 

ChatGPT: A program made by the company OpenAI that gained prominence in late 2022/early 2023 due to its extensive, human-like answers to prompts. The GPT stands for "generative pre-trained transformer". 

Clickbait: use of rhetorical devices to make a user more likely to click on a video or headline (e.g. sensationalised headlines, direct address). Pejorative term.  

Content farm: also called a "content mill", this is a webpage, Youtube channel or similar that quickly creates a large amount of content and uses SEO and clickbait tactics to drive traffic to their website. These sites make their money from advertising. 

Contract Cheating: Cheating that is paid for; for example, buying essays from websites, or paying others to write a paper for you. 

Dirty data: Compromised data; usually through inefficient, disorganised, or unethical recording practices. It could be outdated, mis-typed, or in extreme cases, deliberately falsified (e.g. inventing scores for an assessment task and putting them into ACS). 

Disciplinary Literacy: teaching the ways in which meaning-making functions in your discipline. This might include how particular words are used (e.g. "find" in mathematics), what constitutes valid knowledge, what constitutes valid expression, and how texts in the discipline are structured. 

Generative AI: AI that generates new content in response to a prompt or request. 

Hallucination (AI): an AI "hallucinates" when it gives a confident response that is not based on reality (or, indeed, the reality of its training data). Sometimes referred to as "confabulation". 

Input: Something provided to a person or program to give them information or request something (e.g. a prompt given to Dall-E).

Jailbreaking: Deliberate actions taken by the user to circumvent safety or other software restrictions, usually by exploiting flaws in the software that allow the user to access functions they shouldn't have access to.  

Large Language Model: A type of program that uses a lot of data to train on and uses that data to predict what answers it should generate. Large language models usually give human-like, sophisticated responses. 

Latent Space: the term used for the theoretical "space" that AI uses to undertake its calculations. 

NSFW: Not Safe For Work. Usually denotes content with some measure of nudity, extreme violence, swearing or other content that would be frowned upon by most workplaces. 

NSFL: Not Safe For Life. Usually denotes content that is extremely distressing and traumatic, such as graphic images of death or distressing NSFW content. It is strongly recommended that you do not open anything described as NSFL. 

Open AI: A US company that conducts AI research. They made ChatGPT and have recived considerable funding support from Microsoft. 

Output: An artefact created in response to an input (e.g. a picture, some text) 

Plugin: A software component that adds a feature to an existing program, allowing for customisation by the user. Examples might include ad-blocking plugins or grammar checkers. 

Prompt tailoring/prompt engineering: Emergent terms for the practise of carefully considering the elements included in a prompt for an AI so as to get a more desirable output. 

Prompt injection: The addition of a (usually malicious) instruction to a prompt given to a generative AI model, such as running particular code, or hidden text presented as part of an image that will elicit an unexpected or problematic response. 

Reddit: A large social media site that is organised into smaller forums focussed on specific topics. Users operate pseudonymously. There is a large range of content on Reddit, some of which is undesirable, NSFW, or NSFL; some of which is very wholesome, useful, or interesting. Use with caution.  

SEO: Search Engine Optimisation, or the practice of improving traffic (visitors) to one's website through being found on search engines. This usually involves deliberate choices in keywords, linking, page structure and the back-end of the website

Stack Overflow: A question and answer site for computer programmers. 

Woke: American political term that has made its way into mainstream discourse. Two distinct meanings depending on context and the political views of the speaker. 1. Positive term: Politically and socially aware, particularly regarding social injustices. 2. Pejorative term: used to indicate that the person or thing called "woke" follows an intolerant left-leaning political ideology.