Erick Brethenoux, research director of artificial intelligence at Gartner, has been in a prime position to witness the explosion of interest in generative AI from businesses around the world since ChatGPT launched in 2022. In fact, he said that now, for the first time, even his 83-year-old mother finally understands what he does.
“In fact, she has been very creative in the way she has been using [generative AI]”, said.
However, companies don't always start with a full understanding of generative AI. In an interview with TechRepublic at the Gartner IT Symposium/Xpo in Australia in September, Brethenoux said there's confusion in the market about the technology, in part due to the language vendors are using.
Common misunderstandings include what AI actually is in the broad sense, versus generative AI, and how AI agents differ from generative AI models. This is causing some organizations to make mistakes in how they try to apply the technology to their business use cases.
Confusion about different types of AI
The sudden surge in interest and media attention around generative AI has led to a lot of confusion, as people equate AI as a whole with generative AI capabilities. Brethenoux highlighted that AI is a much broader discipline, with many other important applications beyond generative AI.
“AI and generative AI are not the same thing,” he explained. “They are not interchangeable.”
As Brethenoux explained, generative AI is a practice under the umbrella of AI, while AI is a broad discipline that has many techniques and practices, including decision intelligence, data science, and generative AI.
SEE: Why Teradata believes generative AI projects risk failing without understanding
An example of confusing market terminology is the widespread use of the acronym AI/ML in the field.
“I hate that acronym because it means AI equals ML. That’s not true,” Brethenoux said. “AI techniques are rule-based systems, optimization techniques, graph technologies, search mechanisms, ambient technology — there are all kinds of AI techniques that have been around forever, for the last five decades.”
Generative AI is used in only 5% of production use cases
Brethenoux said that generative AI currently represents only a small proportion of AI in production.
“It’s 90 percent of the airwaves and 5 percent of the use cases,” he explained.
“That’s basically what I see in production today. Of course, if you count the number of co-pilots out there and say it’s generative AI, then the number is much higher now. But until you see a return on investment in that kind of application, to me, that’s not really a use case. It’s just a feature.”
Meanwhile, Brethenoux noted that other AI technologies continue to be used in a variety of use cases.
“What about the rest of AI? Well, that’s why planes arrive on time, because optimization techniques are used to organize all those crews, passengers, planes, airports, gates, and everything. And good luck doing it without AI. All those systems work because AI is the background today.”
AI agents are confused by static AI models
Gartner highlighted AI as a key strategic technology trend to watch in 2025. However, Brethenoux said customers should avoid confusion about what an AI agent actually is, especially when “vendors are very good at confusing our customers” by saying AI models and AI agents are the same thing.
“They are not even remotely the same thing,” he said. “It is actually very damaging to put them in the same sentence.”
Brethenoux added:
- An AI agent is an active software entity that performs tasks on behalf of someone or something and often acts independently.
- An AI model It is a passive entity created by an algorithm and a data set. While an agent may use models to perform its task, they are not the same thing.
SEE: 9 innovative AI use cases in Australian businesses in 2024
“I think the confusion comes from that combination of building a dynamic system that does something and building a set and library of static assets that can be exploited but don’t do anything in particular,” he explained. “They’re just there until you use them. Agents can use them, but they’re not the same thing.”
Confusion in AI leads to costly mistakes for organizations
Brethenoux said he had seen organizations “make big, costly mistakes” as a result of not understanding AI. Some organizations run into problems when they apply a static AI model without having the proper infrastructure to make it dynamic, causing costly delays and other problems in production.
Brethenoux said some confusion was apparent at the Gartner Symposium: “I just had a conversation with a gentleman who was saying, ‘We want to use generative AI for this. ’ And I said, ‘Well, what you’re trying to do can be solved with a graphical technique in a much easier, much cheaper, much faster way. ’”
The AI 'break' is over and now we focus on putting it into practice
Following the launch of ChatGPT, the AI field launched headlong into a period of exploring generative AI models. This marked a shift from the previous focus on operationalizing AI and managing the technical debt associated with deploying large-scale AI systems—what Brethenoux called AI engineering.
Brethenoux said that by January 2024, organizations had returned from this “break” and were making AI engineering a top priority again as they sought to effectively deploy new generative AI capabilities.
“Starting in January 2024, it was a sudden thing for us from a research perspective: recess was over and we were back in the classroom,” he explained. “We were asking, ‘How do we make these damn things work?’, ‘How much money do they cost?’, ‘Are they actually useful?’ and ‘Where do we use them?’ AI engineering is back.”
