𝟭. 𝗜𝗱𝗲𝗮𝘁𝗲 𝗡𝗼𝘁 𝗩𝗮𝗹𝗶𝗱𝗮𝘁𝗲
Use large language models for ideation and creation, but always apply your own knowledge, research, and critical thinking to validate the ideas generated.

Large language models, like ChatGPT, generate responses based on patterns identified from the massive datasets they were trained on and are incredibly powerful tools for creation. However, they can be prone to 'hallucinations/confabulations'—generating plausible-sounding but incorrect or nonsensical information.

Significantly, while they were trained on a large amount of information, they are typically not designed to verify or validate the truthfulness or accuracy of the information they generate and could respond with outdated or even incorrect information. Even as ChatGPT and similar models evolve and become increasingly sophisticated in interacting with information on the internet, our role in discerning and critically analysing their responses remains crucial.

𝟮. 𝗧𝗵𝗶𝗻𝗸 𝗖𝗹𝗲𝗮𝗿𝗹𝘆, 𝗖𝗼𝗺𝗺𝘂𝗻𝗶𝗰𝗮𝘁𝗲 𝗖𝗹𝗲𝗮𝗿𝗹𝘆
Be clear in what you want the LLM to do, and use the right words to achieve your intent.

We may marvel at the responses generated by LLMs, but it's important to understand that they generate content based on the clarity and specificity of the input they receive.

The oft-cited adage 'garbage in, garbage out' applies very aptly to LLMs. If the input is vague, the output may be unhelpful or off the mark. The more explicit and precise we are in framing our thoughts into words, the more specific and useful the responses tend to be.

𝟯. 𝗞𝗻𝗼𝘄 𝗬𝗼𝘂𝗿 𝗦𝘁𝘂𝗳𝗳
Large language models are tools that can enhance our knowledge and understanding, but they don't replace the need for our own expertise and continued learning.

While their responses can seem incredibly accurate and informative, they're not infallible and can generate incorrect or misleading information. If anything, this makes it even more important to have a solid foundation of your own knowledge and not to rely solely on these models for information or decision-making.

Consider them as additional tools in your arsenal, not the ultimate source of truth. Maintain your expertise, stay updated in your field, and use these models as a supplement, not a substitute.


Keep these 3 broad principles in mind when using ChatGPT, Bard, Claude, or any other LLM that you may come across. LLMs are a means to enhance human intelligence and creativity, not to replace them. Use the capabilities of large language models to complement your own expertise, critical thinking, and up-to-date knowledge, and you can harness their full potential while minimising risks.