A collection of some AI tools that are available to use for free (or which have trial accounts). Check the terms and conditions carefully before you use them. Some require a Google account to access. If you try these tools, think about what forms of ML are being used, and how it is being processed.
Also consider the ethical and academic integrity implications of your experiments. Don't pass them off as your own work.
Image: "Robot With A Pearl Earring", created in StableDiffusion via PlaygroundAI.
Once you've gone beyond testing a few tools out, it is well worth experimenting in these spaces to see how they are built, and how the coding works. Free to sign up, and you can really learn a lot from the people and spaces here.
There are many demonstrations (free or commercial) of Generative AI tools that you could try. It is best to describe these are generative, rather than creative, as they use their training data to generate new outputs, but are derivative of the data put into them. These data might be language (large language models) or images (diffusion models). AI tools use natural language processing to try to interpret your prompt (instructions) to generate their outputs.
These tools use Natural Language Processing to turn user-created prompts into text outputs, based on Large Language Models. Their use is already common in SEO-optimised social media and content for advertising. In recent news, OpenAI's ChatGPT has been very popular, though is not easily accessible here in China.
"Academic integrity is a principle in education and a choice to act in a responsible way so others can trust us. It means conducting all aspects of your academic life in a responsible and ethical manner. The IB expects students to produce genuine and authentic pieces of work, that represent their own abilities." IB. 2022.
GPT3 text might be easy to generate, and can be very useful, but it can also be detected. if you are using AI text solutions, you still must be aware of academic integrity and think critically about the work it is producing.
Try some queries in the search box below, or go to the bottom of the page to see loads of examples.
These tools take user prompts and turn them into images using a Latent Diffusion Model. Some of the most popular image creators at the moment are DALL-E 2 (by OpenAI), Stable Diffusion (by StabilityAI) and MidJourney. Avoid paid mobile apps - some are making profit from models trained on artists' work. Consider the ethical and safe use of these tools before using.
At WAB we have a DeepRacer club and activities during some Day 9 blocks. This is an opportunity for students in Middle School and above to try the concepts of reinforcement learning models as they develop models to train agents.
Explore the site and test it out - what can it create for you?