What are the different types of AI? A practical guide

Understanding AI types by deployment category — the governance questions each raises and where documented failures cluster.


AGC
AI Generated, Human Reviewed

Most guides to AI types focus on technical architecture — narrow vs general, reactive vs limited memory. This one focuses on deployment categories: the AI systems actually in use, the governance questions they raise, and where documented failures tend to cluster.

Each card links to a dedicated page covering the specific risk profile, EU AI Act relevance, and Ethics Score links for tools in that category.

Looking for a specific tool? The AI Hub scores 169 tools individually across six governance dimensions. Documented failures are in the Cases library.



Type 01

Conversational AI

Chatbots, virtual assistants and LLM interfaces. Governance concerns: hallucination, identity deception, jailbreaking and data retention.

Read more →


Type 02

Writing Assistants

AI content generation, editing and copywriting tools. Governance concerns: training data copyright, attribution failure and misinformation production.

Read more →


Type 03

AI in Finance

Credit scoring, fraud detection and algorithmic trading. EU AI Act high-risk classification. Bias and explainability are the central governance concerns.

Read more →


Type 04

Customer Service Bots

Automated customer interaction at scale. Governance concerns: escalation failure, vulnerable user exposure and complaint mishandling.

Read more →


Type 05

AI in Education

Adaptive learning platforms and automated assessment. EU AI Act high-risk. Minor data privacy and algorithmic bias in grading are documented concerns.

Read more →


Type 06

Image and Video AI

Generative image and video, deepfakes and synthetic media. Non-consensual intimate imagery and political disinformation are the primary documented harms.

Read more →


Type 07

Speech AI

Voice recognition, text-to-speech and voice cloning. Biometric data handling and fraud enablement via cloned voices are the primary governance gaps.

Read more →


Type 08

Marketing Automation AI

AI-driven targeting, profiling and campaign optimisation. Dark patterns, subliminal targeting and consent erosion are the documented concerns.

Read more →


Type 09

Retail and E-Commerce AI

Recommendation engines, dynamic pricing and inventory AI. Algorithmic pricing discrimination and consumer profiling are the central governance issues.

Read more →


Type 10

Personal Assistants

Always-on voice and chat assistants. Ambient data collection, cross-service data sharing and psychological dependency are documented concerns.

Read more →


Type 11

Creative AI

Music, art and design generation tools. Training data copyright disputes and output attribution are unresolved at law in most jurisdictions.

Read more →


Type 12

AI for Gaming

NPC behaviour, matchmaking and monetisation AI. Engagement manipulation and behavioural profiling of minors are the primary player protection concerns.

Read more →


QUESTIONS

What are the different types of AI?

The different types of AI can be categorised by capability (narrow AI vs general AI) or by deployment context. By deployment — which is more useful for governance — the main types are conversational AI, writing assistants, AI in finance, customer service bots, educational AI, image and video AI, speech AI, marketing automation, retail AI, personal assistants, creative AI, and gaming AI. Each carries a different risk profile and regulatory exposure.

What is the most common type of AI in use today?

The most widely deployed AI type is narrow AI — systems trained to perform specific tasks. Within that, recommendation engines and conversational AI are the most pervasive by user contact volume. Most people interact with AI dozens of times daily through search, navigation, spam filters, and social feeds — often without recognising it.

Which type of AI is highest-risk under the EU AI Act?

The EU AI Act explicitly classifies AI in finance (credit scoring), AI in education, and AI in employment as high-risk under Annex III. These require conformity assessments and human oversight before deployment. Prohibited practices — including subliminal manipulation and CSAM — apply across all categories. The Frameworks section covers how the regulatory alignment dimension is scored.

Where can I find ethical reviews of AI tools by type?

The AI Hub covers 169 tools scored across six dimensions: transparency, data privacy, safety architecture, corporate conduct, bias mitigation, and regulatory alignment. Documented governance failures by AI type are in the Cases library.