SocietyAI glossary
Prompt injection
An attack that involves hiding malicious instructions in content an AI will read (web page, email, document): the AI takes them for legitimate commands and executes them. Unlike jailbreaking, the attacker does not speak directly to the model but traps the data, and this has become the number one security flaw in AI agents.
In other words
Slipping a fake instruction signed "the boss" into the mail pile of a highly obedient assistant, who carries it out without a second thought.
Alyriaz articles that mention it
- Claude Sonnet 5: Anthropic sells (almost) Opus-level power at a standard model price
Anthropic launched Claude Sonnet 5 on 30 June 2026, the default model for Free and Pro. It comes close to Opus 4.8 on agentic tasks while costing 60% less during the launch period. We break down the figures, the strategy, and what to expect from it day to day.
- Agentjacking: the flaw that turns your code assistant against you
An attack dubbed Agentjacking, disclosed in June 2026, traps AI coding agents by injecting fake instructions into error reports. 85% success rate, 2,388 organisations affected. We explain how it works and how to protect against it.