# Recent History of AI (2022–2025)

The last few years have been a distinct era: generative models, public safety debates, and regulation. This sub-article expands [Section 7 of History](../history.md) and complements [Recent Topics](../recent.md).

## The ChatGPT Era (2022)
- **ChatGPT** (Nov 2022) brought LLMs to the public.
- Adoption exploded; companies and institutions raced to integrate.

## Frontier Models and Multimodal Systems (2023–2025)
- GPT-4, Claude, Gemini, Llama, and open models advanced rapidly.
- Multimodal and agentic systems emerged. See [The Generative and Agentic Era](generative-agentic-era.md).

## Safety and Governance Milestones
- The 2023 "pause" letter and Geoffrey Hinton's warnings. See [The AI Safety Debate](ai-safety-debate.md).
- **Bletchley Summit** (2023) and **EU AI Act** (2024).
- Compute governance and export controls expanded. See [Compute Governance](compute-governance.md).

## Why It Matters
This era shifted AI from research to infrastructure, making safety, trust, and governance central.

## Sources
- OpenAI. *ChatGPT announcement* (2022).
- European Parliament. *AI Act* (2024).
- Bletchley Declaration (2023).

*Back to [History](../history.md) · See [Recent Topics](../recent.md).*