Best AI tips for beginners
Local
When Catholic Education SA held its Communications and Marketing Conference in Adelaide during April, one of the workshops focused on artificial intelligence (AI) obedience training. Delegates enjoyed learning about AI during a workshop with marketing specialist and creative director Steve Davis, who runs Talked About Marketing with business partner David Olney.
The duo’s AI Obedience School helps people master the art of working with and prompting tools like ChatGPT and Perplexity.
Here are Steve’s top five tips for AI beginners.
- Define your objectives
Before diving into AI, know exactly what you want to achieve. For instance, if you’re looking to draft a blog post, specify the topic, tone and key points you want to cover. This clarity helps in selecting the right AI tool and setting precise parameters for your task.
- Craft clear, direct instructions
AI tools thrive on specificity. For example, when asking ChatGPT to draft a blog post, instead of saying ‘write something about AI undermining human creativity’, a more effective prompt would be similar to the one I used during my presentation at the conference. I picked up on some cues from one of our attendees who voiced their fears about AI undermining human creativity.
After some discussion with the participant and some quick research into what other thinkers were saying on the topic, I used the following ‘prompt’ to ask ChatGPT to create a first draft. ‘My name is James, I am a Catholic school teacher and I am worried about the impact of AI on my students, namely, diminishing the development of their own robust creativity. Please draft a blog post that alludes to my fear that this development is like all the worst bits of Blade Runner and The Matrix, brought forward to today in the sphere of education. Here are some background notes.’ I then pasted in a few hundred words of research.
- Use specialised tools for research and summarisation
The research alluded to previously, was carried out using my favourite tool, Perplexity AI. This ‘answer engine’ asks you for a topic to research and then scours the web for relevant sources.
When ready, it produces the list of its cited sources and drafts a bespoke summary of the findings. You can continue asking further questions or simply review the summary.
Ideally, you would follow the links to the cited sources to check that they are bona fide, or at least reflect on the summary to ensure it resonates well with your understanding of the topic.
- Engage a human for final checks
AI output should not be the final step. For example, after generating a draft using ChatGPT, have a human review it for nuances, accuracy and alignment with your ‘brand voice’. This step is crucial for quality control and maintaining a personal touch because when left to its own devices ChatGPT can be very wordy and fluffy.
- The power of politeness in AI communication
In a recent episode of our podcast, Talking About Marketing, we discussed observations made by my colleague, David Olney, former lecturer at the University of Adelaide. He found that engaging AI with courtesy, such as by framing requests to ChatGPT with ‘please’ and ‘thank you’, can enhance the quality of responses. For instance, saying ‘could you please help me understand the basics of quantum computing?’ is more likely to yield a comprehensive and considerate reply than a blunt command. This practice not only reflects on the potential of respectful communication to improve interactions with AI but also emphasises the broader importance of kindness in all forms of communication.
To listen to the free podcast or to book a workshop go to talkedaboutmarketing.com