What is the AI?

#1 слайд

1 слайд

#2 слайд
What Is Generative AI? These systems follow programmed instructions and use data to improve accuracy over time. Artificial intelligence (AI) refers to computer-based systems designed to process information, recognise patterns and complete specific tasks automatically. AI technologies can analyse language, process information and generate responses based on predefined rules and patterns. Generative AI is a type of machine learning where computer systems analyse huge amounts of data and then generate new content, such as text, images or videos, based on instructions such as a text prompt.

2 слайд

#3 слайд
How Does Generative AI Work? Generative AI starts with an input; this is some sort of prompt. A prompt could be  an image, text, a design or even musical notes. Generative AI models use neural networks to identify structures and patterns within data. AI algorithms then generate new content to the user in response to the prompt. These outputs can be in the same form as the input or completely different, such as text-to-text or text-to-image.

3 слайд

#4 слайд
History of Generative AI 1957 A book was published which described the grammatical rules for generating natural language sentences. 1932 A machine was invented that could translate between languages using a mechanical computer. This was an early step toward language processing in AI technologies. 1964 - 1966 The first chatbot called ELIZA was created. It was the first program that simulated conversation with the user.

4 слайд

#5 слайд
History of Generative AI 1980 A video game called Rogue was developed. It used early generative AI techniques to create new, randomised game levels each time it was played. This made the gameplay different with every session and influenced many future games. 1968-1970 SHRDLU was developed. One of the first AI systems capable of understanding and responding to natural language in a virtual environment. It could interpret typed commands and manipulate objects in a simulated world, making it an early example of a multimodal AI system. 2006 ImageNet, a large image database, was created. It became a key resource for training AI systems in visual object recognition and laid the groundwork for major advances in computer vision technologies.

5 слайд

#6 слайд
History of Generative AI 2020 An advanced artificial neural network called GPT-3 was released. It was designed to process and generate human language by identifying patterns in large amounts of text data. 2011 A voice-activated personal assistant called Siri was released on smartphones. It could generate spoken responses and carry out tasks based on user voice commands. 2023 An advanced artificial neural network was developed that could process both images and text.

6 слайд

#7 слайд
Different Types of Generative AI Language Models Language models are used for a wide range of tasks, such as translation, academic writing, code writing, grammar correction and even genetic sequencing. Audio and Speech Models Audio and speech models are able to develop and edit songs, support dictation and transcription, create accompanying sounds or noises for video footage, add dubbing to videos and recognise speech and voice. Visual and Imagery Models Visual and imagery models are used to generate 3D images and models, graphs, avatars, illustrations and videos. These models can also be used to edit and enhance existing images.

7 слайд

#8 слайд
Advantages of Generative AI Generative AI can be used to generate or improve new content, such as text, videos and images, that can be hard to distinguish from human-created content. Generative AI can be used to analyse large amounts of complex data quickly and in new ways. This is useful for businesses and science research. Generative AI can help to save time and resources by automating and accelerating a range of tasks and processes.

8 слайд

#9 слайд
Disadvantages and Limitations of Generative AI Businesses may need to invest large amounts of money to incorporate, use, develop and run generative AI into their working practices. The quality of generated content can vary. AI models can generate errors and are unable to filter out bias, personal views or inappropriate content. Users of generative AI can face difficulties in copyright and privacy issues. Generative AI models are trained on large amounts of data, which makes it difficult to check if generated content violates copyright or privacy laws.

9 слайд

#10 слайд
What Can You Remember? Using small pieces of paper or sticky notes, work with a partner to record what you have learnt about Generative AI. Write each fact or piece of information on a separate sticky note or piece of paper. Share your information with the rest of the class.

10 слайд

#11 слайд

11 слайд