Editor’s note: This article was rewritten in October 2026. An earlier version presented 2023 trends as current and quoted statistics we could not verify, so they have been removed.
Looking back from 2026, 2023 was the year generative AI moved from demos into everyday tools. Four trends from that year still shape how AI is built and used.
1. Large language models went mainstream
ChatGPT, released at the end of 2022, became the fastest-growing consumer app of its time in early 2023. In March 2023 OpenAI released GPT-4, which handled longer and harder tasks and accepted images as input. Google answered with Bard and, in December, announced its Gemini models. Writing, coding and research tools built on these models spread into offices and classrooms within months.
2. Open-weight models arrived
Meta released Llama 2 in July 2023 with a licence that allowed commercial use, and a wave of open models followed. They let companies run capable models on their own servers, which is the basis of most self-hosted and fine-tuned AI setups today.
3. AI regulation got real
- United States: NIST published its AI Risk Management Framework in January 2023, and the White House issued an executive order on AI safety in October.
- European Union: lawmakers reached political agreement on the AI Act in December 2023. It became the first broad AI law in a major market.
4. Hallucinations became a known risk
As more people relied on chatbots, so did cases of confident but wrong answers, including invented legal citations filed in court. Checking AI output against sources and keeping a person in the loop became standard advice, and it still applies.
What it led to
Most of what followed, including multimodal assistants, AI agents and on-device models in phones, grew out of these 2023 shifts. For current tools, see our AI Tools coverage.
