AI centre

ExploreAI – AI demos showing the art of the possible today

ExploreAI is a site with a collection of practical AI demos, including examples from cloud suppliers such as Microsoft Azure, Amazon Web Services and IBM Watson, along with examples from Jisc where we wanted to show particular concepts.  Our aim is to help people understand the kinds of things that are possible with AI today.

Image of ExploreAI header.


We started working on the site when we noticed that people were either proposing AI do something that in reality isn’t going to be feasible any time soon, or were dismissing AI as a solution when in fact it was perfectly possible. We came up with the idea of a site where people could get hands-on experience of the kinds of things that are possible to help with this discussion.  You can try the site at

How the site works

Many of the demos are based on services from providers such as Microsoft or Amazon, as part of their cloud-based web services. We’ve done that as these services can be used today by any developer, without needing specialist data science expertise. There are a few exceptions, where we wanted to show a particular educational concept.

We’ve grouped the demos into three areas – Vision, Language and Knowledge.

An view of Explore AI showing the main sections: All, Vision, Language and Knowledge.

Each demo is structured in the same way – we provide some instructions, some sample data, and sections with more information, which provide some technical background, and ‘Things to consider’, which might be legal or ethical issues, or technical limitations.

Screen shot of sentiment analysis

You can either select some of the sample content or provide your own content, and click the action button (in this case ‘Analyse’) to send the content to one or more solution providers, and then see the results.  In this case, we have sent it to Azure, Amazon Web Services, and IBM Watson.

Showing the senitment tool in action

The reason we sent it to multiple service providers is to allow you to explore the difference between different providers – for example, in the sentiment analysis example above, some providers analyse the data on a sentence basis, some on the whole text, and some look for key phrases.

Getting Started

If you are new to AI, here some ways that you might like to get started using the site:

Sentiment Analysis – is the sentiment behind text positive, negative or neutral?

Sentiment analysis image

This is a good place to start exploring, as the concept is familiar to many people.

  • Have a look at the results from the ‘Essay feedback example’.  Do they look right to you? Could they be used as part of an educational process, for example in part of a service to aggregate or highlight feedback to students?
  • Paste some of your own text in – make sure it’s about a paragraph so it has enough to work with. Do the results look right? Can you find some of the limitations?

Face Detection – locating faces in images, and creating information about those faces.

Image representing the face detection section.

This is an interesting example to explore if you want to start exploring AI and vision, and perhaps starting thinking about some ethical issues.

  • Try the example with the demo content and maybe a picture of yourself, and see if you can understand the difference between face recognition and face detection.
  • The service has a guess at your age – this is built into the Azure and AWS service.  Are you surprised by that? What legal and ethical issues does it raise? Is it accurate?

Question generation – using AI to create question and answer sets from passages of text.

This one is a little different in that it’s based directly on work by researchers, rather than using a cloud service.  It’s an example of AI generating, or supporting the generation, of educational content.

  • Can you see any value in this? For example, students using it themselves to test their understanding of something they had read, or helping tutors generate formative assessment questions.
  • How good are the questions?
  • How good are the answers? It’s a harder problem so they often aren’t perfect. And if they aren’t good, does that matter – ie could just the question generation be useful.

More ideas

If you have questions or ideas about ExploreAI then contact us at

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.