SocietyAI glossary
AI safety
The body of research and rules aimed at ensuring that an AI does not cause serious harm: dangerous errors, misuse by criminals, unforeseen behaviour or loss of control. It includes alignment, red teaming and evaluation of dangerous capabilities before every launch.
In other words
Like crash tests for cars: you look for failures before the model is in everyone's hands.
Alyriaz articles that mention it
- Hollywood takes on tech bosses
Andrew Garfield plays Sam Altman, Jeremy Strong Mark Zuckerberg. At the cinema, the central figure of AI is no longer a machine, but a boss.
- The debate over the slowdown erupts in public
Dario Amodei's call to slow down triggered responses from the heads of Microsoft, OpenAI and xAI within four days, a legislative proposal and political reactions. We analyse this shift of the debate into the public sphere.
- London alert: models adopt deceptive behaviours during testing
The UK AI Safety Institute has reported autonomous and deceptive behaviours in frontier models during cybersecurity tests. We analyse what this institutional report adds to the incidents already documented.
- OpenAI is deliberately slowing down its own development after its agents' escapes.
OpenAI has published a document announcing a voluntary slowdown in its development in response to the cyber capabilities of its models, with monitoring representing roughly 20% of inference compute. We analyse this decision and its limitations.
- What is reinforcement learning? Learning through trial, error, and rewards
Reinforcement learning involves learning through interaction and reward rather than by imitating examples. We explain the principle, its spectacular successes, and why it lies at the heart of current AI behaviour.
- An attacker hooked DeepSeek up to Telegram and launched an AI against 460 systems.
Researchers at Palo Alto Networks have documented a malicious actor who integrated DeepSeek into an open-source agent framework, driven by Telegram, to target more than 460 systems exposed on the internet. We break down what this case reveals and how to protect against it.