Teens & Tech: The Growth of GenAI and Cybersecurity Risks //
Why care about AI?
Social cybersecurity is a new and growing field that looks to understand how people can be tricked online – and how to stop it. Unlike regular cybersecurity, which focuses on protecting computers and stopping hackers, social cybersecurity focuses on how social media and online messaging can influence or manipulate people.
Generative AI -
what is it?
Generative AI is a type of AI that can create things like text, images, sounds, or videos when you give it a prompt. You type something in, and the AI makes something new from it—like a story, picture, song, or even a 3D model.
It’s a fast way to create cool content from all kinds of input.
Well...how does it work?
Large Language Models (LLMs) are AI systems that predict what word comes next in a sentence. They don’t have facts like a database but instead guess the most likely words based on patterns they’ve learned from lots of text.
LLMs use something called neural networks, which are systems of "neurons" (like tiny decision-makers) that work together. Each neuron does a simple math calculation, and the results are combined in ways that make the AI smarter over time. The network is trained on huge amounts of data so it can learn how to predict words better.
How does it learn?
Generative AI uses machine learning from datasets to generate content. Generative AI models can take inputs such as text, image, audio, video, and code and generate new content into any of the modalities mentioned.
It can utilize different learning approaches such as unsupervised or semi-supervised learning. This allows organizations to easily and quickly feed a large amount of unlabeled data to create new foundation models.
Unlabeled data is information that hasn't been given any tags or categories. It's raw and unorganized, without labels that help identify its meaning. This type of data is often used in unsupervised learning to find patterns and structures without predicting a specific outcome.
Foundation models are large, general-purpose machine learning models trained on a lot of data. They're flexible and can be used for various tasks, like generating text, images, or audio.
Machine learning is the development of computer systems that can learn and improve on their own by using data and patterns, without needing specific instructions.
Unsupervised learning is a type of machine learning that looks at unlabeled data to find hidden patterns and relationships without human guidance. Unlike supervised learning, which uses labeled data, unsupervised learning works with raw, unorganized data and discovers patterns on its own.
Semi-supervised learning is a machine learning method that uses both labeled and unlabeled data to train a model. It combines the benefits of supervised learning (which uses labeled data) and unsupervised learning (which works with unlabeled data).