AI on a par with the Internet and smartphones

Ever since the U.S. company OpenAI took the world by storm with ChatGPT, market observers agree that a world-changing technology is emerging. As a reminder: After the release in November 2022, it took five days for the number of users worldwide to exceed one million. In January 2023, the figure was already over 100 million. Never before has a technology spread so quickly across the globe.

ChatGPT is what is known as a large language model (LLM). Huge volumes of data and the computing power of state-of-the-art mainframe computers went into its development. That’s because it takes many petabytes of text data from books, articles, or software code for a generative AI to be able to recognize generalizable grammatical and semantic patterns and rules, and integrate them into its language system. To be able to communicate in a human-like manner or generate code, LLMs must be extremely variable and internalize the systematic choice between a host of parameters in their independent training. “The number of these parameters is an order of magnitude larger than the number of stars in our galaxy,” as Marcel Franke, AI expert at Microsoft Germany, recently explained at the third “AI For Laser Technology” conference in Aachen. Despite this complexity, he said that AI was child’s play and intuitive to use because, for the first time, it is no longer humans who need to learn the language of the machine, but the machine that can now communicate in human language. “Because our language serves as an interface, AI becomes an intelligent co-pilot,” he explained. At the same time, he said, large language models could be used in a variety of ways and were suitable for specific applications in industry, administration, the legal sector, or in research and development. “Humans stay in the loop and set the direction, while AI assists,” said Franke, describing the future division of tasks.

Machines speak our language and become partners

His assessment is in line with analyses by the U.S. market research company Gartner. Its “Gartner Hype Cycle” has long become a catchphrase that conjures up the typical progression of innovations: first the steep rise of the hype, followed by a fall into the deep valley of disillusionment, before the tough climb back to market success begins.

Gartner analysts Mary Mesaglio and Don Scheibenreif also classify AI in this cycle, with the peak of the hype curve currently being reached, as they see it. And yet a lot of things are different, they say: “AI is no normal technology trend. It will change our lives as profoundly as the Word Wide Web has done since 1993 and smartphones since 2007.” They, too, see the use of human language as a paradigm shift that raises the partnership and collaboration between human being and machine to a new level. By 2030, they predict that 80 percent of all people worldwide will interact daily with smart robots, and that by 2025, AI will contribute to the development of every third new drug and material. And in 2024, they expect three quarters of all companies to raise their AI investments.

Franke and his Microsoft colleague Ansgar Heinen reported in Aachen on huge development momentum. They said that AI was spreading rapidly in industry and accelerating automation: whether it’s flowchart analyses, software code or the analysis of camera and sensor data in all-round process control. For a year now, the use of LLM had ensured an unprecedented pace of innovation worldwide. “Current market analyses predict annual growth rates of 50 percent for AI in the industrial environment,” they reported, saying that this assessment coincided with their experience. “I have never before in my IT career experienced such a rapid adaptation of a young technology,” reported Heinen, saying that this applied not only to typical first movers, but also to more conservative German companies in mechanical engineering, the chemical and pharmaceutical industry, and automotive engineering. He said that adaptation was taking place at a rapid pace and not even stopping at very advanced development projects: Mercedes, for example, had equipped the new E-Class with voice-controlled AI almost on the fly. Chemical companies are using the models to improve their formulations and manufacturing processes. Elsewhere, intuitive voice interfaces are used to link processes more intelligently or improve cross-departmental collaboration. If a problem arose in production, the employees would be able to describe it in their own language. AI structures and then analyzes these reports and forwards them to the teams in service, design, and development with specific indications of possible causes. “AI is a game changer,” he explained, especially with the return on investment sometimes being achieved within just a few months.