BRAND IDENTITY
The Six-Legged Dog: Eni's brand identity and history
Please, fill this field
Or , our new artificial intelligence tool.
Company
Governance
Sustainability
Governance Go to
Vision
Actions
Products
Vision Go to
Actions Go to
Energy Diversification
Technologies for the transition
Technologies for the transition Go to
Partnership for innovation
Activities around the world
Investors
Media
Careers
- TECHNOLOGIES AND DEVELOPMENT
What is artificial intelligence, how it works and what it’s for
A clear guide to AI: what it is, how it works, the main types, where it’s used, and the benefits and risks in life and work.
COLUMN: The words of energy
What is artificial intelligence
Artificial intelligence (AI) is a computer system’s ability to analyse large amounts of data, identify patterns and trends, and learn from them to make predictions or generate content. Some systems can adapt their behaviour over time, but always under human oversight. Common examples include software that recognises objects in images, virtual assistants and automated writing tools.
How AI works
AI rests on two pillars: data and algorithms. Data are the raw material, while algorithms are the engine that processes information to recognise patterns, predict outcomes or create new content. Machine Learning (ML) combines these concepts: it is a set of algorithms that learn from data rather than being hand-coded case by case.
As in the human brain, AI also uses neural networks: layers of algorithms that process data, identify relationships and produce an output. When there are many layers, this is called Deep Learning, a class of machine learning that uses deep neural networks to autonomously extract complex features and correlations from data. Deep learning allows systems to work on raw, unstructured data, learn the most relevant features autonomously and provide users with more advanced analytical capabilities.
Types of artificial intelligence
AI is commonly distinguished between:
- Narrow AI: it specialises in a single task, such as spam filters or recommendation systems. It is the most widespread form today.
- General AI: it is still experimental and aims for broader, flexible understanding, comparable to human intelligence.
Language models (LLMs) and generative AI
Large Language Models (LLMs) are neural networks trained on vast amounts of texts to learn language patterns and generate new content on request. They underpin generative AI, which today can create text, images, audio and video. Due to their power, they require ongoing human review and responsible data use.
What artificial intelligence is for
AI has by now become part of everyday life: it helps with online searches, film or music suggestions, automatic translation and image recognition - recommendation and computer-vision systems that personalise experiences and save time. In industry and research it is applied across many fields: medical diagnostics, predictive maintenance, process optimisation and the analysis of complex datasets.
Benefits and risks of AI
AI offers significant benefits: it boosts efficiency, automates repetitive tasks and improves decision quality. It can support sustainability by optimising processes and resources. However, there are challenges to manage: data bias, model opacity and privacy protection. That is why human oversight (human-in-the-loop) is always needed to ensure accuracy, eliminate cognitive biases, verify correctness and take responsibility for content in the context where it is presented.
Column: the words of energy
A series of articles that explain in simple terms the energy that powers the world - from natural gas to renewables, from sustainable mobility to decarbonisation.
Download document
