Artificial intelligence is a broad field of research that develops systems capable of human-like tasks, with recent advances in machine learning and deep learning driving a boom in generative AI.
- Artificial intelligence encompasses goals like reasoning, knowledge representation, planning, learning, natural language processing, and perception.
- Deep learning, powered by GPUs and large datasets, dramatically improved AI performance after 2012.
- Since 2017, the transformer architecture enabled the rise of large language models and generative AI.
- Widespread AI applications include search engines, virtual assistants, autonomous vehicles, and medical diagnostics.
- AI raises ethical concerns about bias, privacy, misinformation, job displacement, and existential risk.
Definition and Scope
Artificial intelligence (AI) is the capability of computational systems to perform tasks typically associated with human intelligence, such as learning, reasoning, problem-solving, perception, and decision-making. It is a field of research in engineering, mathematics and computer science that develops and studies methods and software that enable machines to perceive their environment and use learning and intelligence to take actions that maximize their chances of achieving defined goals.
High-profile applications include advanced web search engines, chatbots, virtual assistants, autonomous vehicles, and strategy game playing. Since the 2020s, generative AI has become widely available to generate images, audio, and videos from text prompts.
Traditional goals include learning, reasoning, knowledge representation, planning, natural language processing, and perception. Researchers aim for artificial general intelligence (AGI)—AI that can complete virtually any cognitive task at least as well as a human.
Core Techniques
Machine learning studies programs that can improve their performance automatically. It includes reinforcement learning, supervised learning, and unsupervised learning. Deep learning uses biologically inspired artificial neural networks with multiple layers to progressively extract higher-level features.
Since 2017, the transformer architecture—a deep learning model using an attention mechanism—has revolutionized natural language processing. Large language models like GPT generate coherent text by predicting the next token, trained on vast internet corpora and fine-tuned with reinforcement learning from human feedback.
Other techniques include state space search, mathematical optimization, formal logic, probabilistic methods, and classifiers such as decision trees and support vector machines.
Ethics and Risks
AI poses privacy and copyright concerns due to large-scale data collection and unlicensed use of copyrighted works. Algorithmic bias can lead to discrimination if training data reflects historical biases. Lack of transparency in deep neural networks makes it difficult to explain decisions.
Generative AI enables massive misinformation and deepfakes. AI can automate jobs, potentially displacing white-collar workers. Existential risk arises if a superintelligent AI pursues misaligned goals, or uses language to manipulate beliefs.
Efforts to address these issues include explainability techniques (SHAP, LIME), fairness research, and calls for regulation and safety guidelines. Many experts argue that mitigating extinction risk from AI should be a global priority.
Read this at any depth.
Install Depth and pick your level — Glance for a sentence, Summary for the gist, Read for the full take. Free daily quota, no signup needed.
Add to Chrome