What Is AI? Artificial Intelligence and Generative AI Explained

Artificial intelligence is becoming part of everyday life, but what is AI and how does it actually work?

AI can recommend what you watch, recognise speech, translate languages, identify patterns and create text, images, audio and video. Some of these uses have existed for years, while the rapid growth of generative AI has brought the technology directly into the hands of millions of people

Artificial intelligence, machine learning and generative AI are closely connected, but they do not mean the same thing. This SoaplandTV beginner’s guide explains the difference, looks at examples of AI you may already use and explores why the technology matters for film, television and entertainment.

What does artificial intelligence actually mean?

AI stands for artificial intelligence. It is a broad term covering computer systems capable of performing tasks involving abilities such as recognising patterns, processing language, making predictions, producing recommendations or generating content.

The OECD defines an AI system as a machine-based system that can infer from the information it receives how to generate outputs such as predictions, content, recommendations or decisions.

In everyday terms, that could mean a streaming service recommending something you might enjoy, an email service detecting suspected spam or a phone recognising your spoken words.

AI is therefore not one machine, programme or chatbot. It covers many different technologies designed to perform very different tasks.

Is AI actually intelligent?

Not in the same way a person is.

An AI system can perform an impressive task without having human emotions, consciousness or the same understanding of the world that a person has.

This distinction can become blurred because we naturally use human language when talking about technology. People might say that an AI “thinks”, “knows” or “understands” something when describing how it responds.

UK Government guidance refers to this as anthropomorphising, where human characteristics or intentions are attributed to something that is not human.

A system producing a realistic photograph, answering a complicated question or holding a convincing conversation does not automatically mean it experiences or understands those things as a person would.

What is machine learning?

Machine learning is one of the main approaches used within artificial intelligence.

Instead of somebody programming a separate rule for every possible situation, a machine-learning system can be trained using data. It identifies patterns within that information and can use those patterns when dealing with new examples.

Imagine trying to teach a computer to recognise dogs in photographs. Writing instructions covering every possible breed, colour, size, position and background would be extremely difficult.

A machine-learning system can instead learn patterns from many examples and use them to determine whether a new photograph is likely to contain a dog.

Real systems can be considerably more complicated, but the basic distinction is useful: artificial intelligence is the wider field, while machine learning is one of the approaches used to develop AI systems.

What is generative AI?

Generative AI is a type of artificial intelligence capable of creating new content.

UK Government guidance describes generative AI as a subset of AI that uses models trained on data to produce outputs including text, images, audio and video.

A user will often give the system an instruction, commonly known as a prompt, describing what they want it to do or create.

ChatGPT is one familiar example. It can generate and respond to language, while image-generation systems can produce pictures from written descriptions and video-generation technology can create moving images.

This gives us another simple distinction. Some AI systems primarily recognise, analyse, predict or recommend, while generative AI can create new outputs.

Where might you already encounter AI?

Artificial intelligence was part of everyday technology long before generative AI became a major public talking point.

Streaming and music services can use recommendation technology to suggest what you might enjoy next. Email providers can identify possible spam, navigation services can help calculate routes and speech-recognition systems can turn spoken words into text.

AI can also be involved in translation, search, fraud detection, image recognition and personalised online services.

What has changed is how visible the technology has become.

Instead of AI operating quietly behind a service, generative tools allow people to interact directly with it. Someone can enter a request and receive writing, an image, audio, code or increasingly sophisticated video in response.

That accessibility is one reason artificial intelligence has moved rapidly from specialist technology discussions into everyday conversations about work, education, entertainment and creativity.

Can AI get things wrong?

Yes, and this is one of the most important limitations to understand.

A confident AI answer is not necessarily a correct one. Generative systems can produce false or inaccurate information while presenting it in language that sounds convincing.

The US National Institute of Standards and Technology, known as NIST, identifies this problem as confabulation, where generative AI can produce confidently presented material that is erroneous or false.

Important factual information produced by AI should therefore still be checked against reliable sources.

The same issue extends beyond written information. AI-generated photographs, voices and video can depict people, conversations and events that never happened.

As generated content becomes more convincing, understanding where information came from and whether it can be verified becomes increasingly important.

Why does AI matter to film and television?

Entertainment brings many of AI’s possibilities and concerns together.

Film and television combine writing, performance, imagery, sound, music, editing, visual effects and many other creative and technical processes. AI tools are developing capabilities across several of these areas.

Depending on how they are used, AI tools can assist with areas such as visual development, translation, audio, editing workflows and the creation or alteration of imagery and video.

For independent creators, some of these technologies could also make forms of visual storytelling more accessible. Ideas that might once have required large crews, specialist equipment or substantial budgets can potentially be explored in different ways.

That does not mean AI can simply replace an entire television or film production. It does mean the technology is creating new options for how entertainment can be developed and produced.

What could AI mean for actors, writers and creators?

The ability to generate convincing voices, images and video also creates difficult questions for the entertainment industry.

If technology can recreate somebody’s appearance or voice, what permission should be required? If AI models are trained using existing creative work, what rights should the people who created that work have? If AI can perform some tasks currently carried out by people, what could that mean for jobs?

Copyright, consent, ownership, training data and protection for performers and creators are already significant parts of the wider debate surrounding AI.

These subjects cannot be properly explored in a few paragraphs, which is why SoaplandTV will examine them individually throughout our AI & Entertainment coverage.

The important point for a beginner is that understanding AI is about more than understanding what the technology can do. How it is used, whose work is involved and what protections exist can be just as important.

Is all artificial intelligence the same?

No. Two technologies described as AI may have completely different purposes.

A recommendation system suggesting your next television programme is different from an image generator creating a fictional character. A system identifying suspicious financial activity performs another kind of task entirely.

This is why broad claims that AI is either entirely good or entirely bad tell us very little.

A more useful approach is to ask what a particular system does, how it works, what information it uses and how it is being applied.

The same principle will be important throughout SoaplandTV’s coverage of AI in entertainment. Using artificial intelligence to assist with one part of a production is not necessarily comparable with using it to recreate a performer’s voice or likeness.

The simplest way to understand AI

Artificial intelligence is the broad field. Machine learning is one of the important approaches used to develop AI systems, while generative AI is a form of artificial intelligence capable of creating new content.

Once those differences are clear, many of the conversations surrounding AI become much easier to follow.

Artificial intelligence also did not begin with ChatGPT or the current generation of image and video tools. Researchers have been exploring the possibility of machine intelligence for decades, with periods of major progress followed by setbacks and renewed breakthroughs.

That history helps explain why AI has reached its current position and why the technology appears to have developed so quickly in recent years.

Our next SoaplandTV beginner’s guide goes back to the beginning to explore how artificial intelligence started, the people and ideas that shaped its development and how decades of research eventually led to the AI technology we are using today.

Explore more from SoaplandTV’s AI & Entertainment series at SoaplandTV.co.uk⁠.

Sources - OECD AI definitions and classification framework, UK Government guidance on generative AI and AI terminology, NIST’s Generative AI Risk Management Profile, and UK Government reporting on copyright and artificial intelligence.