Democratization of IT — as AI moves into management
AI has gone from technical curiosity to topping the agenda in boardrooms in a short time. Despite headlines about “new industrial revolution,” we're not here to spread the hype -- quite the opposite. In a sober but far-sighted spirit, it can be stated that AI is already rewriting the rules of business logic. Companies that have traditionally been stable see their business models challenged by algorithms and self-learning systems. In short: it is no longer an issue if AI will affect business, without how.
For today's business leaders, it is therefore a matter of understanding the technology in depth in order to be able to guide development — not just letting the IT department keep the wheel. Sure, AI's rampage may seem dizzying, but sticking your head in the sand is about as effective as ignoring the elephant in the server room. Business logic is changing at a supersonic pace, and far-sighted organizations are already preparing. This review comes at the right time—to provide a credible, strategic look at where AI is taking us next, without falling into the limp and with the twinkle in the eye.
Read more on a larger screen
AI and LLM — from the ivory tower to everyone's toolbox
No IT revolution in modern times has been as democratic as AI's entry into operations. Technologies that previously required PhDs or dedicated IT departments are now within reach of anyone with a web browser. Back in 2016, Microsoft declared that it wanted to “take AI down from the ivory towers and make it accessible to everyone,” and today we're seeing just that happen in real time. Generative AI services such as ChatGPT — the most popular major language model (LLM) for consumers — spread explosively in 2023, reaching over 200 million active users. One can safely say that AI is quite talked about nowadays (understatement of the year) and well on its way to becoming as mundane as a cup of tea in the afternoon.
To put it in context: Artificial intelligence (AI) is a collective term for technology in which computers can perform tasks that normally require human intelligence -- everything from recognizing patterns to making decisions. Large language models (LLMs) are a branch of AI trained on huge amounts of text to understand and generate human language. OpenAI's GPT-3, with its 175 billion parameters, demonstrated the leap in scalability and capability of these models, and the successor GPT-4 takes it even further with multimodal skills (it can handle both text and image).
LLMs are already being used practically for customer service chats, automated document analysis, code generation and creative tasks such as writing draft marketing materials. This is where the democratization of AI takes place: tools that previously required language and AI experts are now available in the form of user-friendly services integrated into everyday applications.
The future looks interesting to say the least. Swedish initiatives such as AI Sweden, supported by Vinnova, are developing their own large-scale language models for Swedish and Nordic languages — partly to reduce dependence on foreign tech giants and secure digital self-determination. At the same time, a flora of more specialized AI models is predicted to emerge. A recent Gartner report predicts that smaller, tailor-made language models for specific domains could be used three times more than generic giant models by 2027. In other words, we are moving towards a future where AI is not only widely available, but also fine-tuned for every conceivable need — all under the premise that power over technology is shifted closer to the business.