🤖 What is artificial intelligence?
Computers that learn from examples instead of fixed rules.
Artificial intelligence (AI) is when computers learn to do smart things, like understanding speech, recognizing pictures or answering questions. They learn by looking at lots and lots of examples, a bit like you learn by practicing.
AI is the part of computer science that builds systems able to do tasks that normally need human intelligence. Most modern AI uses machine learning: instead of being programmed rule by rule, a model learns patterns from large amounts of data. Deep learning uses artificial neural networks with many layers and powers image recognition, translation and chatbots. Large language models learn from huge amounts of text to predict the next word, which lets them write and answer questions. AI is very useful but can make mistakes or repeat biases found in its training data.
Modern AI is dominated by deep learning: large parameterized models trained by gradient descent. Transformers, built on self-attention, underpin large language models pretrained on next-token prediction and then adapted with supervised fine-tuning and reinforcement learning from human or AI feedback. Capabilities improve with compute, data and parameters, with diminishing returns and behaviors that are hard to predict. Open problems include hallucination, robustness, interpretability, evaluation and alignment with human intent. Inference cost, data provenance and energy use shape real-world deployment.
⚡ Quick facts
- The term "artificial intelligence" was coined in 1956, at the Dartmouth workshop.
- In 1997, IBM's Deep Blue beat world chess champion Garry Kasparov.
- Transformers, the design behind today's chatbots, were introduced in 2017.