Generative artificial intelligence (GenAI) has passed the peak of Gartner’s hype cycle but has not lived up to expectations, analysts warned at the company’s European conference in Barcelona.
In her opening keynote at the European Gartner Symposium, Alicia Mullery, vice president of research at the analyst firm, discussed two AI races. The second is to deliver the results of AI safely and reliably. “This is your race,” she told the audience of IT executives.
One of the takeaways from the opening session is that it’s easy to waste money with GenAI. “You need to understand this bill and keep an eye on it,” warned Mallery and co-presenter Darryl Plummer, Gartner principal research analyst.
Plummer noted that the majority of organizations Gartner spoke to are not ready for AI. “They’re not emotionally, technically, organizationally or managerially ready for it,” he says.
To minimize failures. Gartner recommends two approaches. One is primarily for organizations looking to use AI to improve productivity. The second focused on using AI to drive transformation.
According to Gartner data, running a proof-of-concept project can cost anywhere from $300,000 to well over $2 million. IT and business leaders may be aware of the significant costs associated with training AI models on expensive graphics processing unit (GPU) hardware, but the costs associated with AI inference quickly It can get out of hand, Plummer said.
“AI models are very expensive to process because they have to use something called matrix multiplication to process all the parameters used to get the predictions, which requires a GPU. You buy GPUs and install them in your own data center, or you lease them from a cloud provider, both of which are very expensive,” he said.
Plummer warned that technology providers are too focused on viewing AI advances from their own perspective and not taking their customers along on the journey to achieve the goals of these advanced AI systems. “Microsoft, Google, Amazon, Oracle, Meta, and Open AI have made one big mistake: They’re showing us what we can do, but they’re not showing us what we should do. ” he said.
Plummer says many organizations aren’t ready to adopt the advanced AI offered by large providers, so 75% of their budget goes toward IT consulting to understand how new technology can benefit their organization. He said he was aware of the amount spent on
“We need more budget to get to the proof-of-concept stage,” he said, adding that costs will continue to rise until IT leaders start deploying enterprise AI systems into production environments. Manage ongoing costs.
Analysts explained that IT leaders need to consider the outcomes they want to achieve. Companies looking to implement AI to achieve business efficiency improvements (what Gartner calls “AI Stable” organizations) are likely to have fewer than 10 pilots or AI initiatives in place. There is. In this scenario, you can delegate monitoring and checking to ensure that the AI system is working properly.
Organizations that see GenAI as an industry-changing technology are likely to conduct more pilots. Gartner classifies these organizations as “AI-powered organizations.” The analyst firm believes that humans cannot control the AI systems that AI-powered organizations are considering deploying.
The company therefore foresees the rise of a technology called TRiSM (Trust, Risk, and Security Management), which it said will play a key role in ensuring compliance of AI systems.
