Wake The Town
Gear Seven/Arc Studios/Shift
Contemplative Reptile
  • International Edition
  • USA Edition
  • UK Edition
  • Australian Edition
  • Canadian Edition
  • Irish Edition
  • German Edition
  • French Edition
  • Singapore Edition
  • Spanish edition
  • Polish edition
  • Indian Edition
  • Middle East edition
  • South African Edition

Don’t Be the Last to Try AI!


Veronica Millan, global chief information officer at MullenLowe Group, and our resident web3 expert for Little Black Book, encourages the industry to quit talking about AI and begin using it

Don’t Be the Last to Try AI!

Five months ago, a creative director in our agency was talking to me about AI and how being on the precipice of this technology is taking us into the next chapter of creative advertising. So, I had to ask him if he – personally – ever used any of these products? Did he have an account with Midjourney or DALL-E 2? Had he tried an internet search using Microsoft Bing powered with ChatGPT to see what kind of conversation you can have with it? 

It's not enough to talk about. You have to try it out because AI is not just ‘coming’ to the advertising industry, it is already here. And you can’t really appreciate what the use of these AI tools can mean to your organisation until you have experienced it for yourself. Most of our industry has been exploring these AI tools for the last several months to better understand the possibilities of AI and/or show clients what can be accomplished in this interplay of advertising and next-generation technology. We can’t afford to ignore this tech; it’s changing how we work.

As an example of how you could use these AI tools, if you want a brief on a social media campaign for a new beverage, your AI engine can start with knowledge of other winning or effective work that’s been done before in that brand’s space. Someone else in one of our agencies was playing with ChatGPT and was amazed at the brief it generated using the parameters he had entered into the tool. The consensus was that this was not anything you could ever present to the client as a final product, but it provided a strong foundation from which to work. Compare these AI-generated starting points with relying solely on the knowledge and research capabilities of the people in the room.  

And now, there are a few AI tools that are targeting the marketing and advertising space. These tools are trained to consider the specific language and context of advertising as well as our particular industry knowledge. In theory, this will allow these apps to generate better answers for advertising agencies than a general, publicly available tool like ChatGPT or Midjourney. These AI tools are taking marketing-specific inputs, including proprietary information (like case studies from previous, award-winning work) and adding that data into their artificial intelligence models. 

These models are intelligent in the sense that they can quickly call up these concepts and put them together. But as anyone who has played around with AI-generated images or copy has seen, the horses can have an extra leg or the copy can sound really wrong to the ear. While AI tools can process a great deal of data quickly, they're not very ‘smart’. They can’t always tell when they have misrepresented how an animal looks or generated copy that sounds a little too much like Shakespeare and not enough like current English. In the end, AI tools are machine learning, or large language, models; they are not magical or sentient beings that can think and reason. This is why even Bing, as a well-developed and well-funded search engine, is making headlines about the weird conversations it has with people. It’s still a machine, taking in and feeding back the information that we input, rather than coming up with new knowledge. In other words, artificial intelligence still needs humans to make it have sense.

The same is true for AI tools used in the advertising industry; they don’t eliminate the work that has to be done. We still need to get approvals of the final concept, put it into production, hire talent, licence artwork, create the artwork, photography, and so on. The AI doesn’t do the final work. But there is something about AI assisting in the process that feels both incredibly helpful and democratic: anyone can ask it to design something with the right questions, rather than asking a design expert to mock something up for you.

But the bottom line is that you should be considering and using this tech. Explore it, see the limits of it, understand how it’s working and decide what role it’s going to play in your work. As we move into this new era of technology-assisted work, we need to know what we are using, so we can start to ask it to do better work, and decide where it adds the most value to our business. 

And next month [in this column], we’ll talk about the dark side of this technology.

[And no, this was not written by ChatGPT or any other AI tool. I promise!]


Veronica Millan is global chief information officer at MullenLowe Group. She writes a regular column about the metaverse for Little Black Book - check out the rest of the series here

view more - Thought Leaders
Sign up to our newsletters and stay up to date with the best work and breaking ad news from around the world.
MullenLowe Group, Mon, 13 Mar 2023 11:33:31 GMT