Skip to main content

What we can do better than AI

By September 4, 2023November 1st, 20232 Comments

We love ChatGPT. Period. Our communication team hardly ever have a writers-block moment since they use ChatGPT. We are now scrolling on twitter, not for the scandal-binging or collection of weird memes (well, we still do that), but to save and hoard threads like these:

Clients often ask us: “Aren’t you guys worried about ChatGPT and AI taking over your communication projects?” And for a while, we did worry. Because with just a couple of prompts, AI can produce that social media caption, press release, one-pager, infographic, feasibility report, script, article, recipe, manual, case studies, and what have you not in a matter of seconds.

Why, then, would clients stick with us and still use our services for IEC (Information, Education, Communication) materials?

Because there are a couple things we still do better than AI:

1- We know what the client needs to convey and what the audience wants to hear. With the right prompts, AI can provide a plethora of options in disseminating your message. But can AI also discern and select the right combination out of all these options that would be most effective, relevant, and compelling for your audience? No matter how good the prompt, AI still needs us humans to formulate the right mix

2-We have that special sauce. OpenAI can generate a well structured piece of text with flawless grammar and that is great if all you need is a well structured piece of text with flawless grammar. OpenAI can identify algorithms from company documents you feed them and develop materials using a similar tone and manner. But what we would like to call the soul of the company, capturing the energy, aspirations, humour, and approach of the people behind it, is still embedded in the writer’s ability to absorb the environment of the client and reflect that in its writing

We think of OpenAI and ChatGPT as pattern makers that save tailors time in the time consuming process of drawing. But as most seasoned tailors do, we review the generated patterns, tweak and adjust based on what we know about the customer, and design according to the wishes of the customer and think of ways to make the result even more sophisticated and comfortable for the wearer. Ultimately, that is still something only a human can do.

Join the discussion 2 Comments

  • rio says:

    AI dikembangkan dengan pembelajaran mesin (Machine Learning). Semakin sering orang menggunakan ChatGPT, semakin banyak dia mempelajari pola pikir manusia lalu membangun model yang semakin serupa dengan prilaku manusia .. sampai pada titik tertentu (mungkin) bedanya tipis dengan manusia 🙂