AGI? GPUs? Learn the definitions of the most common AI terms to enter our vocabulary

In today's fast-evolving technological landscape, artificial intelligence (AI) has become an integral part of our lives, influencing everything from banking to healthcare. However, understanding the terminology associated with AI can feel daunting, especially as industry leaders and policymakers often communicate in a jargon-filled language that can be difficult to decipher. To help you engage with the world of AI, here’s a comprehensive glossary of key terms and concepts you should be familiar with: **Agentic AI** refers to systems capable of making autonomous decisions with minimal human intervention, distinguishing it from reactive models like ChatGPT, which require prompts to operate. **Artificial General Intelligence (AGI)** denotes AI that can perform complex cognitive tasks akin to human reasoning, encompassing self-awareness and critical thinking. **Alignment** is a critical area of AI safety research focused on ensuring that AI systems operate in accordance with human values and objectives. **Bias** in AI emerges from the data these models are trained on, often reflecting human prejudices and inaccuracies, potentially leading to skewed results. **Capability Overhang** describes the disparity between the potential of AI models and the extent to which their applications are currently utilized. This term was popularized by Microsoft’s CTO, Kevin Scott. **ChatGPT**, the flagship chatbot developed by OpenAI, launched with significant excitement in 2022, has since sparked a competitive race in the AI sector, with rivals continuously striving to surpass its capabilities. **Claude**, Anthropic’s main AI model released in March 2023, has garnered attention for its coding abilities, and enhancements in healthcare applications were introduced in 2026. **Compute Resources** are essential for training AI models, encompassing GPUs and cloud services that facilitate extensive data processing. **Data Centers** house powerful computing hardware designed for managing vast amounts of information necessary for AI functionality. **Deepfake** technology creates realistic digital content through AI, raising ethical concerns due to its potential for misuse in deception. **Distillation** involves transferring learned knowledge from a larger AI model to a smaller one, simplifying the replication of AI technologies. **Doomers** are skeptics who warn against the possible dangers of unchecked AI development, fearing it could pose risks to humanity. In the context of AI, **Effective Altruists** advocate for using AI to address significant social challenges like climate change and poverty. **Federal Preemption** highlights the ongoing debate regarding whether AI policies should be set at the state level or centralized federally, with significant implications for regulation. **Frontier Models** represent the cutting-edge of AI technology, capable of performing a variety of tasks that surpass existing capabilities. **Gemini**, relaunched by Google in 2023, has been positioned as a strong competitor to ChatGPT, with updates suggesting it meets or exceeds prior benchmarks. In terms of energy needs, **Gigawatts** measure the substantial power requirements for operating AI infrastructures, making it a critical consideration for industry leaders. **GPUs** are specialized chips essential for training AI models, with companies like Nvidia playing a pivotal role in providing the necessary technology. **Hallucinations** refer to instances when AI generates incorrect information as if it were factual, a phenomenon that raises questions about reliability. **Large Language Models (LLMs)**, such as OpenAI's GPT-5 and Google's Gemini, are designed to generate human-like text responses based on extensive datasets. **Machine Learning** enables AI to learn independently over time, continuously improving without explicit programming. **Multimodal AI** can interpret and generate outputs across various formats including text, images, and audio, enhancing user interaction. **Natural Language Processing (NLP)** encompasses techniques for understanding and generating human language, crucial for AI communication. **Neural Networks** mimic human brain processes, used in applications like facial recognition to improve accuracy. **Open-source** refers to software accessible for public use and modification, with calls for transparency in AI training processes. **Prompt Engineering** is the art of crafting questions to elicit desired responses from AI systems, increasingly recognized as a specialized skill. **Rationalists** advocate for a logical, evidence-based approach to understanding AI's potential, opposing those who rely on intuition or emotion. **Responsible Scaling Policies** are critical for ensuring the ethical development of AI, balancing innovation with societal impacts. **Singularity** is a theoretical point when AI surpasses human intelligence, often depicted in science fiction. **Transformers** are foundational to the architecture of LLMs like GPT, enabling efficient data processing and model training. **Universal Basic Income (UBI)** has gained attention as a potential safety net in an AI-driven economy, reflecting fears of job displacement. **Key Figures** in the AI landscape include prominent leaders such as Sam Altman of OpenAI, Elon Musk of xAI, and Demis Hassabis of Google DeepMind, all of whom are shaping the future of this transformative technology. Understanding these terms will not only enhance your grasp of artificial intelligence but also prepare you for the discussions shaping our collective future.