162 articles
Browse guides
Deep learning stacks many layers of simple mathematical units so that each layer learns a slightly more abstract representation than the one below it. Depth is what lets a model go from pixels to edges to shapes to ‘cat’ without anyone defining what an edge is. Every large language model and image generator in use today is deep learning. In practice: A vision model’s first layers detect edges; its last layers detect faces — nobody programmed either step. ...
Deepfakes combine generative video or voice cloning with a real person’s likeness. Detection is unreliable and getting worse, so the practical defences are provenance signals, verification channels, and organisational process. The EU AI Act imposes disclosure obligations on this kind of content. In practice: A cloned CEO voice on a call authorising a wire transfer.
If your company uses an AI system at work, you are probably a deployer. Purely personal, non-professional use is excluded. Deployer duties are lighter than provider duties but real for high-risk systems: use it as instructed, assign competent human oversight, keep logs, and inform affected people. In practice: An HR team using a third-party CV screening tool is a deployer of a high-risk system. Plain-English summary of Article 3(4). The binding text is Regulation (EU) 2024/1689. ...
Differential privacy adds calibrated noise so that individual records cannot be reverse-engineered from outputs, with a tunable privacy budget quantifying the guarantee. Unlike anonymisation, it is a proof rather than a hope. The cost is accuracy, and the budget is finite across queries. In practice: Publishing aggregate statistics that cannot be used to identify any single respondent.
A diffusion model is trained to reverse a noising process. Given pure static and a text prompt, it iteratively denoises until an image emerges that matches the prompt. It replaced GANs as the default because it trains more stably and scales better. In practice: Fifty denoising steps between random static and a finished illustration.
In distillation a large teacher model generates outputs that train a smaller student. The student ends up cheaper and faster while retaining much of the teacher’s behaviour on the target distribution. Most affordable production models are distilled from something larger. In practice: A model 10× smaller and 20× cheaper that handles 90% of your traffic just as well.
Downstream providers build products on top of another party’s GPAI model. The Act obliges the upstream model provider to supply the technical information downstream providers need to meet their own obligations. It is the mechanism that makes the value chain legible instead of a black box. In practice: Building and selling a legal assistant on top of someone else’s API. Plain-English summary of Article 3(68). The binding text is Regulation (EU) 2024/1689. ...
An embedding turns text, an image, or an item into a vector in a high-dimensional space arranged so that semantic similarity becomes geometric proximity. That single property powers semantic search, recommendations, clustering, and the retrieval half of RAG. In practice: ‘car’ and ‘automobile’ land near each other; ‘car’ and ‘carpet’ do not, despite sharing letters.
Article 3(39) defines it and Article 5 prohibits its use at work and in education, with narrow exceptions for medical and safety purposes. Elsewhere it is generally high-risk. The scientific basis for inferring emotion from faces or voice is contested, which is part of why the restriction is strict. In practice: Software claiming to score candidate enthusiasm from a video interview. Plain-English summary of Article 3(39) and Article 5. The binding text is Regulation (EU) 2024/1689. ...
Training normally takes many epochs, since one look at the data is rarely enough. Too few and the model underfits; too many and it starts memorising. Very large models sometimes train for less than a single epoch simply because the dataset is enormous. In practice: Ten epochs over 50,000 images means the model saw each image ten times.