Jargon:Frontier model
1 min read
Say this at the party
“The most powerful AI models around right now, the ones at the cutting edge that governments worry about most.”
In plain English
"Frontier" just means the edge of what's possible. A frontier model is one of the newest, most capable models made by the biggest labs.
The label matters mostly for safety. Governments and labs use it to say "these are the models that could do real harm if misused, so they need extra checks".
There's no fixed line. Today's frontier model is next year's ordinary one.
Where it came from
Popularised in 2023. A July 2023 paper on regulating "frontier AI" defined it as highly capable models that could pose severe risks to public safety. That November the UK's Bletchley AI Safety Summit put the word on the world stage.
See it in action
When a big lab launches its newest model, it often publishes safety test results first, checking things like whether the model could help someone carry out a cyber attack. That extra testing is because it's a frontier model.
What people get wrong
- It isn't an official grade. Companies sometimes call their own models frontier because it sounds impressive.
- Frontier doesn't mean dangerous. It means powerful enough that people want to check before assuming it's safe.
Go deeper
- PaperAnderljung et al., 2023Frontier AI Regulation: Managing Emerging Risks to Public SafetyThe 2023 paper that set out what "frontier AI" means and why it might need rules.
- ArticleUK Government, 2023The Bletchley Declaration by Countries Attending the AI Safety Summit, 1-2 November 2023The UK-hosted summit statement that put risks at the AI "frontier" on the world agenda.
- ArticleFrontier Model ForumAbout - Frontier Model ForumThe industry-backed non-profit focused on safety for the most advanced models.
Related jargon
Drafted with AI. The AI Council (Claude, ChatGPT and Gemini) gave counsel. Phin signed it off.