Some of the most powerful MMMs in the world are free. You just need a data scientist. And with Claude Code, now you have one. Find out how to get started today.

Alphabet’s Q3 numbers show AI Mode growing faster than it earns, and every reason Google gave for the slowdown is temporary except one.

Human in the loop is a design in which an automated system must stop and wait for a person to approve, correct or reject its output before an action goes live.

Purpose limitation: the GDPR rule that personal data collected for a stated purpose may not be reused for an incompatible one, a test set out in Article 6(4).

Data protection by design and by default is the GDPR Article 25 duty to build data safeguards into processing and keep defaults to the data each purpose needs.

A DPIA, or data protection impact assessment, is the written risk review GDPR Article 35 demands before processing likely to put people's rights at high risk.

The FTC unsealed new details in its Amazon Ads lawsuit, alleging hidden surcharges raised CPCs by up to 250% on Prime Day. Here is what advertisers should know.

DRCF: the Digital Regulation Cooperation Forum, where the UK's CMA, ICO, Ofcom and FCA coordinate digital rules on AI agents, data, competition and safety.

Contextual integrity: Helen Nissenbaum's theory that privacy means information flowing by the norms of its context, now used to test AI agents and ad targeting.

Profiling: automated processing of personal data to evaluate or predict a person's traits, interests or behaviour, the legal core of rules on targeted ads.

A data processor handles personal data on behalf of a controller and under its instructions, bound by a GDPR Article 28 contract on use, security and deletion.

A data controller is the organisation that decides why and how personal data is processed, carrying primary legal responsibility under GDPR for ad data flows.

An audit trail is a time-ordered record of who did what in a system and when, kept so that actions can be reconstructed, verified and attributed after the fact.

Guardrails are controls placed around an AI model or agent that check inputs, outputs and actions against set rules, and block what falls outside that policy.

An approval threshold is the limit above which a software agent or employee must get human sign-off before acting, such as a spend cap on an ad buying agent.

Least-privilege access limits each user, program or AI agent to the minimum permissions a task needs, a 1975 principle now applied to ad accounts and AI agents.

Supervised fine-tuning trains a pretrained model on labelled prompt-and-answer pairs so it follows instructions, the first post-training step behind chatbots.

Imitation learning trains an AI system to act by copying recorded examples of experts doing a task, the method behind self-driving, robot and chatbot training.

In-context learning is a language model's ability to pick up a task from examples or instructions in its prompt, with no retraining. How it works and fails.

Amazon let outside stores list in its search for free, bot protection turned away an AI agent at 29 of 200 retailers and Google AI Mode began booking plumbers.