Both Traditional AI and Generative AI (Gen AI)are based on machine learning – algorithms that enable machines to learn from data without being explicitly programmed. However they differ in important ways.

Traditional AI

Traditional AI models are mostly focused on advanced analytic tasks such as prediction or clustering. They typically work with just one form of data and are trained for one specific task that they become very proficient at. They need to be trained from scratch or ‘fine-tuned’ to perform well on a different task.

Generative AI

Gen AI models on the other hand are based on foundation models that can ingest data of various types (text, image, video) and output any or all of these types. Gen AI models require lots more data (insanely huge amounts) than traditional AI models. But, when they are provided with that volume of data, they are more versatile: they can do things they weren’t explicitly trained for, because the immense volume of data gives them more options to work with.

Neither form of AI is necessarily ‘better’ or ‘worse’ than the other. Each has its own benefits depending on the intended use, and in fact, in many practical situations, the best solution involves a combination of both.mile. Offering ‘meaningfully different’ CX will help supercharge your business performance and make your brand more resilient, especially important in the current consumer and market context.

A spectrum between two modes of exploration

Brilliant minds have been working on the challenge of ‘machines that think’ since the 1940s when the fundamental building blocks first were conceived. Fast forward 80 years and ChatGPT (released 30 November 2022) is ubiquitous and it has been swiftly followed by varied generative brethren.

Undoubtedly far closer than ever to the imaginations of Claude Shannon (‘father of information theory’) and Alan Turing (‘father of theoretical computer science and AI’), the advent of these tools has been described in many quarters as signalling a fundamental shift in how industries, and even the world, must operate.

Undeniably, some of the capabilities that Gen AI implementations have demonstrated are intriguing, entertaining and impressive. Equally, some applications are distinctly troubling, whether for industry, labour or broader society. Therefore, you should consider a spectrum between two modes of exploration: enthusiastic experimentation ‘because we can!’ at one extreme, compared to thoughtful ‘should we, and what could go wrong?’ testing and deployment at the other.

Both modes are valid and rewarding – the path chosen is often based on the values and objectives of the organisation, team or individual investigating the technology. Kantar has structured the development of tools for our business and clients so that the former (enthusiastic experimentation) takes place in the (secured, controlled) back room so as to incubate leaps in capability. Then, we can systematically build and deliver the latter (thoughtfully tested) solutions with confidence.

Ultimately, it’s all about striving for the balance of enthusiasm and care when applying Gen AI (and indeed, any other technology) to our approaches, methods and techniques – and by extension to the goals and objectives of the brands we partner with.

John Cucka
Head of Analytics,
Kantar Australia
john.cucka@kantar.com

Rob Kramer
AI Lead, Analytics
Kantar Australia
rob.kramer@kantar.com