[DigitalToday reporter Jinju Hong (홍진주)] As generative artificial intelligence (AI) spreads rapidly from finance and healthcare to streaming and the auto industry, understanding key terms used inside and outside the tech sector is becoming a basic requirement for reading industry trends. As the AI race begins to shake up infrastructure, regulation and labour-market change beyond simple service launches, industry terminology is increasingly becoming a common language across the wider economy.

In this regard, Business Insider on May 25 (local time) compiled key concepts, companies and figures that explain the generative AI race.

One of the most closely watched ideas in the AI industry recently is “agentic AI”. It refers to AI that makes its own judgments and carries out sequential tasks with only limited human involvement. It is seen as a sign that generative AI has moved beyond simple question-and-answer since the arrival of ChatGPT and is shifting toward actually performing work. The industry’s long-term target is AGI, or artificial general intelligence, which refers to AI that carries out complex cognitive tasks like humans.

The core foundation of the AI race ultimately is computing resources and infrastructure. Graphics processing units (GPUs) have become key chips used to train and run inference for AI models, and demand is also surging for AI data centres that can house such large-scale chips, storage devices and power facilities. In particular, the latest AI data centres require far greater power and cooling facilities than earlier generations. According to the outlet, 1 gigawatt can supply power to about 750,000 households, and 10 gigawatts corresponds to computing power equivalent to about 4 million to 5 million GPUs. It means competition in the AI industry is spreading into a race to secure electricity, semiconductors and networks.

Key terms that determine model performance include large language models (LLMs), transformers and context windows. An LLM is a program that learns vast amounts of data to generate human-like sentences. Transformers, a core part of that structure, cut training time and enabled larger models by processing large amounts of data simultaneously rather than sequentially. A context window refers to the range a model can remember in a conversation. A wider range helps with understanding longer contexts and reducing hallucinations.

Safety and trust are also key variables in the AI industry. Alignment is a field of safety research that seeks to match AI goals and behaviour with human values and intent, while bias refers to the problem of AI absorbing human discrimination and errors in training data as they are. Hallucinations are cases where non-existent information is presented as fact, and deepfakes link to the misuse of AI-generated fake images, video and audio for crime and fraud.

In corporate strategy and the competitive landscape, OpenAI, Anthropic, Google, Meta and Nvidia were cited as central pillars. ChatGPT became the starting point of the generative AI race after its 2022 launch, and Anthropic’s Claude and Google’s Gemini are also increasing their presence in the enterprise market and the competition for high-performance models. Nvidia has emerged as the biggest beneficiary of the surge in demand for AI chips.

Key figures were also treated as an important axis for understanding the AI industry. Sam Altman is a co-founder and chief executive officer of OpenAI, and after being ousted by the board in 2023 and returning within days, he became a leading executive in the AI industry. Dario Amodei, as Anthropic CEO, has repeatedly warned that society is not prepared for the large-scale replacement of jobs that AI could bring. Jensen Huang has led Nvidia and emerged as the biggest beneficiary of the sharp rise in demand for AI chips. Sundar Pichai faced criticism that Google was slow to respond immediately after the arrival of ChatGPT, but later received assessments that Google had narrowed the gap.

Policy disputes around the AI industry are also growing. In the United States, a key issue has emerged over whether to set regulations by state or create a single federal standard. As worries grow that AI could replace jobs, debate over universal basic income is resurfacing. Optimistic views are also pushing back, including those like Elon Musk who say AI will raise productivity and expand prosperity over the long term.

In workplaces, terms such as tokens, multimodal, prompt engineering and open source are also spreading quickly. Recently, expressions have emerged such as “token maxing”, aimed at maximising productivity by increasing AI usage itself, and “vibe coding”, in which AI writes code with little human involvement.

Ultimately, the AI industry is spreading beyond a simple chatbot race into a major industrial reshaping in which computing infrastructure and power, regulation, labour markets and safety issues are intertwined. As AI has entered everyday services broadly, analysis says understanding terminology itself is becoming a basic capability not only for reading the tech sector but also for tracking changes in financial markets and policy.