Download our brochure
Request a demo
en
fr en
Return
Understanding AI generation for recruitment : the technical basics you need to know

Understanding AI generation for recruitment: the technical basics you need to know

Previous article
Next article

What is generative AI ?

Generative AI represents a fascinating facet of artificial intelligence, revolving around learning from data to generate autonomous and original content. This field uses sophisticated automatic data processing techniques to produce works that emulate human creations.

More specifically, generative AI is a specialized branch of AI, exploiting neural network models, also known as NPL (Neuro-Linguistic Programming). Unlike classical predictive models, which rely on existing data to anticipate future outcomes, generative models innovate by creating new data based on pre-existing models.
 

How does generative AI work ?

Generative AI uses trained algorithms composed of deep learning mathematical models and artificial neural networks. These artificial neural networks will be able to design output content from input data. Generative AI learns through observation and repetition.

 
Download infography
 

Example of generative AI

ChatGPT

The ChatGPT conversation tool mainly generates text responses to formulated requests/prompts. It is available in several languages, including French, and incorporates the ability to memorize previous requests within the same conversation.
 
To date, ChatGPT has demonstrated an ability to identify its own errors and deal with inappropriate requests. ChatGPT excels at writing or summarizing texts, drafting e-mails in corporate style, detecting errors in lines of code, solving calculations, as well as creating stories.
 

AI HubSpot

HubSpot provides its users with a free artificial intelligence text generator. This tool generates high-quality content that can be used for blog posts, landing page content or e-mail campaigns.

DALL-E 2

Open AI features an AI-based image generator. This tool is capable of generating images in response to text queries, performing realistic and targeted retouching on images, as well as creating variations from existing visuals.

Gamma

Gamma is a powerful slide generator, for submitting text requests such as questions, instructions or task descriptions.
 
The tool also lets you adjust visuals and content to suit user needs. Options include various graphic lines, adding sounds or videos, and customizing chapters based on the presentation's topic statement.
 
The application operates on a credit-purchase model, with 400 credits offered on registration. A paid pro version, offering unlimited credits, will soon be available.
 
 

All about prompt engineering

Prompt engineering refers to the practice of formulating specific instructions or queries to train AI-assisted language models. Prompt engineering is essential for the use of Artificial Intelligence. Prompt engineers use precise, optimized formulations to achieve desired results, often involving playing with the structure, tone and detail of the query to influence the response generated.

 

How to write prompts properly ?

Start with an action verb

To formulate an effective prompt, start with an action verb. This tells the AI your intention and what you want the result to be.

Feed your query with context

It's vital to feed your query with context so that the AI can understand the purpose of your request. You can include additional information on the subject, who the content is intended for, objectives, expertise and any other relevant elements.

Opt to use references

Giving examples and references helps AI to understand the context. References can be images, videos, documents or any other type of content.

Use quotation marks

Quotation marks are used to indicate to the AI that it should take into account the text between quotation marks, or to consider it as an example, reference or keyword.

Specify answer length

It's a good idea 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 may vary according to the purpose and type of request.

Requesting corrections

Prompt engineering is a process that asks for errors, corrections and requests for adjustments. If you're not happy with the AI results, consider fine-tuning your prompt to achieve the desired results. 

Specify the tone

It's 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 query and your needs. For example, if you want your AI to generate a LinkedIn caption for you, 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 content for a Landing Page, you can ask the AI to think about an outline, then its sub-sections, then finally the content.

Download infography

Conclusion

In summary, generative AI represents an exciting advance in the field of artificial intelligence. Using sophisticated techniques, it excels in the creation of autonomous and original content, clearly distinguishing itself from traditional predictive models by actively generating new data from pre-existing models.

Boost your HR operations with AI superpowers

Request a demo

Stay updated to HRtech trends and news

By signing up to our newsletter

Send
By subscribing to the newsletter, you agree to receive personalized marketing communications by email, and confirm that you have read the privacy policy.