Understanding AI generation for recruitment: the technical basics to know
What is generative AI?
Generative AI represents a fascinating facet of artificial intelligence, centered on learning from data to generate autonomous and original content.
This field uses sophisticated automated data processing techniques to produce works that emulate human creations. More specifically, generative AI is a specialized branch of AI, leveraging neural network models, also known as NPL (Neuro-Linguistic Programming in English).
Unlike classic predictive models that rely on existing data to anticipate future outcomes, generative models innovate by creating new data based on preexisting models.
How does how AI works generative AI work?
Generative artificial intelligence uses trained algorithms made up of deep learning mathematical models and artificial neural networks. These artificial neural networks can create output content from input data. It should be noted that generative AI learns through observation and repetition.
Examples of AI generative AI
ChatGPT
The ChatGPT conversational tool mainly generates text responses to formulated requests/prompts. It is available in several languages, including French, and includes the ability to remember previous requests during the same conversation.
To date, ChatGPT demonstrates an ability to identify its own errors and handle inappropriate requests. ChatGPT excels at writing or summarizing texts, drafting emails in the company's style, detecting errors in code lines, solving calculations, as well as creating stories.
HubSpot AI
HubSpot provides its users with a free AI-powered text generator. This tool generates high-quality content that can be used both for writing blog posts and for creating landing page content or email campaigns.
DALL-E 2
OpenAI has an AI image generator. This tool can produce images in response to text requests, perform realistic and targeted edits on images, as well as create variations from existing visuals.
Gamma
Gamma is a powerful slide generator, by submitting text prompts such as questions, instructions, or task descriptions.
The tool also makes it possible to adjust visuals and content according to users' needs. Options include various graphic styles, adding sounds or videos, and customizing chapters based on the presentation topic statement.
The application works on a credit-purchase model, with 400 credits offered upon sign-up. A paid pro version, offering unlimited credits, will be available soon.
Everything you need to know about prompt engineering
Prompt engineering, or "formulation engineering" in French, refers to the practice of creating instructions or specific requests to train AI-assisted language models. Prompt engineering is essential for using artificial intelligence.
Prompt Engineers use precise and optimized wording to achieve desired results; this often involves playing with the structure, tone, and details of the request to influence the generated response.
How write well your prompts?
Start with an action verb
To craft an effective prompt, start with an action verb. This tells the AI your intent and what result you want.
Provide your request with context
It is essential to provide your request with context so the AI can understand the goal of your request. You can include additional information about the topic, who the content is intended for, the objectives, the level of expertise, as well as any other relevant element.
Choose to use references
Giving examples and references helps the AI understand the context well. References can be images, videos, documents, or any other type of content.
Use quotation marks
Quotation marks are used to tell the AI that it should take the text between the quotes into account or consider it as an example, reference, or keyword.
Specify the length of the response
It is useful to specify the length of the response in order to determine how many “tokens” are needed to produce the answer. The length of the response can vary depending on the objective and the type of request.
Ask for corrections
Prompt engineering is a process that involves errors, corrections, and adjustment requests. If the AI results do not suit you, consider refining your prompt to get the desired results.
Specify the tone
It is highly recommended to specify the tone to give your response a style. The tone of your content can vary depending on the objective of the request and your needs. For example, if you want your AI to generate a LinkedIn caption, you can specify the desired tone: playful, professional, informative...
Ask the AI to think step by step
Ask the AI to think step by step to get a more complete answer. For example, if you want to write the content of a Landing Page, you can ask the AI to think about an outline, then its subsections, and then the content itself.
In short, generative AI is a fascinating advance in the field of artificial intelligence. By using sophisticated techniques, it excels at creating autonomous and original content, standing out clearly from traditional predictive models by actively generating new data from preexisting models.