Artificial intelligence (AI) will drive other emerging technologies to coordinate in the areas of advanced hardware, edge devices and cybersecurity, benefiting business transformation, says Dell Technologies.
The company envisions quantum computing and generative AI intersecting within five years.
"In 2024, we will see an AI era in which generative AI dialogue will move from theory to practice," said John Roese, global chief technology officer of Dell Technologies, in an online roundtable entitled "Visions: 2024 and Beyond".
Dell believes generative AI will shift from building training infrastructure to inference infrastructure, he said. Generative AI in 2024 will be used at scale and have greater impact in terms of transforming business enterprises, said Mr Roese.
Inferencing is the process of using a trained AI model to generate predictions that make decisions or produce outputs based on input data.
In 2024, AI will be put into real practical use, he said.
"We will have to think about security much more aggressively, where people attempt to steal data-trained models," said Mr Roese. "Developing the security architecture around inference infrastructure will be a critical task next year."
The second big shift is the economic discussion on AI projects will move from the cost of data training to overall cost of operation, he said.
The third change is enterprise moving from a broad AI experiment to a top-down strategic focus on picking the few generative AI projects that can truly transform its businesses, said Mr Roese.
Dell predicts the enterprise generative AI supply chain and ecosystem will improve in 2024, particularly in the semiconductor ecosystem, where there will be a greater choice of graphic processing units.
Meanwhile, availability of both closed- and open-source models and tools to help enterprises implement generative AI will continue to expand, he said.
Dell's third vision is zero trust security will become the centre of the discussion, from the conceptual stage through to becoming a reality, said Mr Roese.
The zero trust approach creates environments in which all devices, people, applications and data are authenticated. Only authorised behaviour can take place, while everything else is prohibited.
The ability to understand an anomaly is intrinsic within the infrastructure. Governments and industries around the world have decided to make this approach a requirement.
He said the fourth vision is the edge platform will emerge. Most AI use will happen where data exists, and most data in the world does not exist in data centres. It exists in sensors, actuators, manufacturing environments and retail stores, said Mr Roese.
Generative AI systems are used in the realm of the probabilistic, which means they try to figure out the correct answer by assessing the degree of probability, he said.
These kinds of predictions and the kind of probabilistic activity carried out via generative AI on a conventional computer are very inefficient, Mr Roese.
However, quantum computers are extremely good at figuring out probabilistic and optimisation problems. As these forms of AI and quantum computing have an affinity, he said he expects the AI and quantum ecosystems to become more intertwined.