What Does ChatGPT Understand?

ChatGPT has become an important and significant tool with artificial intelligence (AI). A fundamental question about any such AI is what is the nature of its 'intelligence'; is it truly intelligent, in the sense that it understands the tasks it is set, or does it merely give an impression of understanding. Furthermore, though, does this matter? That is to say, if ChatGPT provides the answers that the user needs does it matter whether it understands the question?

This talk gives a very brief history of the development of AI which led to products like ChatGPT and then addresses the above question: Does ChatGPT understand anything, and does it matter whether it does or not.

Objectives

It is likely that all students have heard of ChatGPT, but will many of them understand what it is? This talk should give them some perspective about the true nature of this kind of AI, and should be accessible to a wide range of students; no technical knowledge is assumed.

There is some cross-over with The Enigma of Enigma in that they both refer to the work of Alan Turing.

There is a video of this talk on YouTube.

Slides

Accent the Positive

People's voices are a very important part of their identity (as is illustrated in this talk), but what happens when a person has to use a synthetic voice? Until recently the most famous and celebrated user of synthetic speech technology was probably Stephen Hawking. The voice that he used was notoriously mechanical and artificial, but it was identified with him as an individual – to the point that great efforts had to be expended on keeping the voice working. More recently, Lee Ridley (Lost Voice Guy) has become very famous in a very different way. He won Britain's Got Talent, relying on speech technology to help him to present his act as a stand-up comedian.

Lee was a member of the Creative Speech Technology - or CreST - research network, and this talk presents some of the outputs from that network. As illustrated in this talk, Lee's voices demonstrate some of the advances in the technology compred to Hawking's – although he still cannot have the Geordie accent that he would like.

I have given a number of versions of this presentation – with variations in its title. It was originally presented at the York Festival of Ideas, then at a number of locations, including twice at Bolton School. The most recent was at Benton Park School, Leeds.

Objectives

The main message of this presentation is to demonstrate the valuable results that can be generated by a truly multi-disciplinary group, which included not only scientists and technologists, but also artists, writers, musicians and users of the technology. It is to be hoped that it will demonstrate to students who may believe they have little interest in technology

Slides

This is the version for Benton Park.

The Enigma of Enigma

It is generally agreed that the fact that the Allies cracked the Nazi Enigma 'code' shortened World War II and hence saved inummerable lives. That the code (more correctly a cipher) was cracked has become well-known, largely thanks to the film, The Imitation Game. As entertaining and informative as that film is, how the code was broken is not clear. It turns out that the design of the cipher and its technology was exceptional and should have been extremely secure; most of the problems were due to the humans who used the technology; they did not follow the instructions of the designers. While it is often too easy for technologists to 'blame the user', in this case they had a point.

This talk introduces the concept of encryption and then tells some of the (human) history of the cracking of Enigma.

Objectives

If anything encryption is even more important in the modern, internet-dominated world than it was in World War II. Without needing a lot of technical background, this talk explains the basic idea behind encryption and its strengths and weaknesses, while adding a bit of history. There are links to the talk on Staying Safe Online.

Randomness

'Oh, that's a bit random', is the sort of thing that is easy to say, but what really is randomness and how do people cope with it – or not? On the contrary, human survival and social evolution rely on the finding of patterns, sometimes to the extent of seeing patterns in processes which are in fact random. Ironically genetic evolution does rely on truly random mutations. Those mutations rely on subatomic interactions, and they are the only examples of genuine randomness.

Otherwise, processes that are random – or appear to be – do play an important part in our everyday lives (e.g. lotteries). This talk shows the importance of these differences.

Objectives

This talk covers an area of mathematics and technology which are important in our everyday lives in a manner intended to be entertaining. This may open students to ideas which they have probably not contemplated previously.

Staying safe online

There is generally a healthy awareness that the Web is a hazardous environment, that when communicating sensitive information, such as bank details, there is a danger of them being diverted and mis-used. Hence, people are aware of the importance of security measures such as passwords and PINs, and yet many probably do not appreciate the mechanisms behind them; why should I include 'upper- and lower-case characters and at least one special character' in my password?

Technologists have a tendancy to 'blame the user': the system would be entirely secure if only the users would follow the rules and (for instance) choose complex but strong passwords. Yet there are reasone why users will act insecurely, after all they are only human.

This talk leads through some of the mechanisms behind security, what works and where its weaknesses are, and then gives some (grounded) ideas as to how to remain as secure as possible. Security needs to be safe, but it also needs to be usable.

Objectives

The main motivation of this talk is for the student to appreciate how the technology works; they probably already know why it is important. It may persuade some to be more careful about their own on-line security.

Can Blind People Do Mathematics?

Mathematics is generally perceived as a cognitive activity: something you do 'in your head'. From that point-of-view there is no reason to believe that blind people should not be able to do mathematics. However, there are limits to how much maths anyone can hold in their head, but also when two or more mathematicians communicate they generally rely on written, visual notations. This can represent a serious barrier to blind people and hence their access to most STEM subjects.

Two projects MATHS and Lambda have addressed this problem, by turning algebra into synthetic speech and braille. Also soundgraphs can be a most powerful auditory representation.

This talk introduces these technologies and the concepts and ideas behind them.

Objectives

This is not a heavily mathematical talk, but it will highlight some of the importance of the discipline. It should also illustrate how an imaginitive approach to the technology can address novel problems.

Introduction to user-centred design

Traditionally it has often felt as if the user of any technological system has been low in the list of the designers' priorities. Fortunately, this is no longer true as is apparent in the improved usability of modern systems – even quite complex ones. This is largely due to the application of user-centred design methods. The basic idea is to start with the user.

This talk is a brief introduction to some of the ideas and techniques that are applied to make this work, including personas, scenarios and conceptual design

Objectives

On one level this talk is intended to raise students' awareness of the importance of design, and the role of the potential user. At another level it might spark interest in the possibilities of careers in design.

Get Set for Computer Science

This talk is suitable for students in year 11 upwards, who have some idea that they might want to go on to study Computer Science at university.

What is Computer Science and what is it like to study it on an undergraduate programme? The talk covers selecting a course/university and making your application. How do you choose between departments and between universities? Having chosen your target universities, how can you maximize your chance of getting a place?

This talk shares a title with a book of the same name and is a summary of the same material. However, it is not just a chance to promote the book!

Objectives

This talk should help students to think about their options in possibly applying to study Computer Science (and other STEM subjects) at university. For those who are going to make such an application it will give an insight into the process, with the objective of ensuring that they end up on the course that best suits them.

Where is the Science in Computer Science?

This talk is based on a classical problem of Computer Science, The Travelling Salesperson Problem. This seems deceptively simple but turns out to be very complex and is an instance of specially hard problems. An objective of this talk is to show that there is a theoretical basis behind computer programming and that an understanding of the theory has vital practical implications.

Objectives

This talk should broaden students' understanding of the scope of the technology. It addresses the importance of STEM subjects by giving a quick introduction to the fact that an understanding of the science behind computers, expressed mathematically, can lead to better technology.

Slides

Computer Science?

This is a variation on the above talk, Where is the Science in Computer Science?. This one is slightly less technical and more general. Again, though, the main objective is to give students an insight into the value of STEM subjects, but Computer Science particularly.

It draws a conclusion that there is such a discipline as Computer Science, but that a lot of subjects given that label are really a branch of engineering.

Slides