ARTIFICIAL INTELLIGENCE

Artist Artist Bethany Hadfield collaborates with Scientist Jonny Jackson to explore his research training Artificial Intelligence to recognise markers of heart disease

Watch

Artist Bethany Hadfield takes you into a digital environment where blood vessels grow and the heart beats. This piece echoes the synthetic digital veins and arteries created to train artificial intelligence to recognise markers of heart disease.

Bethany is interested in exploring a multidisciplinary approach to art making and the ideas that manifest when art and science assimilate. She believes it is important to explore difference, rather than commonalities, and feels that if we bring together two quite different approaches the outcome will be unpredictable, which will result in creating something new for both artist and scientist.

Play

TRAIN THE MACHINE

Scientist Jonny Jackson invites you to answer the questions and contribute to a growing database that will train a fictional Artificial Intelligence to make decisions independently. The piece highlights the moralistic nature of many of the perceived dangers of, and obstacles to, the adoption of AI technology.

Jonny has been involved in several projects beyond his studies, working with artists, writers, actors, and the public. They continue to remind him of the importance of sharing science with diverse communities to influence, inspire and learn from them.

The Research

Artificial Intelligence is a set of technologies that allow computers to automate human intellectual activities and to learn to perform tasks. It has the potential to transform healthcare by supporting doctors to complete faster and more accurate diagnosis of disease.

To teach AI how to perform a new task requires a huge number of meticulously labelled images known as training data. Sourcing and labelling hundreds of thousands of patient scans is labour intensive, time consuming and can sometimes be an additional risk to patients.

Instead, digital synthetic data can be generated. Using mathematics, the biological and physiological factors of the heart can be mimicked. This allows digital veins and arteries to be created and is currently being used to train AI to recognise markers of heart disease with the hope of improving diagnosis.

Have your say

Share your thoughts to help share the future of healthcare technologies.

Previous responses

Would you share your personal data to help develop an AI for medical reasons?

yes, I think that the evolution of this technology could really help the medical world to progress which would help to save people's lives and increase their quality of life so I would be happy to.

Would you share your personal data to help develop an AI for medical reasons?

Yes. Assuming the data is given anonymously, there is no reason why it shouldn't be used.

Would you trust a doctor using AI to make your medical diagnosis?

yes, I think AI might be smarter than a human at spotting trends and abnormalities. I think this has already been proven to be so with identifying breast cancer in mammograms.

Would you trust a doctor using AI to make your medical diagnosis?

Yes, like I would trust my doctor to use other technologies like ultrasounds - I would want to be presented with the results that lead to their conclusion. I’d like to know what limitations there are to the method.

Would you trust a doctor using AI to make your medical diagnosis?

I might trust the doctor to use AI if they had explained the medical ethics and checks and balances in place. Used well it could provide positive tool to advance healthcare. Used badly it could be negative.

Would you share your personal data to help develop an AI for medical reasons?

Depends on who had access to it, but I would be open to the idea

Would you share your personal data to help develop an AI for medical reasons?

Yes, I would, because I think it is our moral duty to co-operate for the good of the whole community.

Would you trust a doctor using AI to make your medical diagnosis?

Yes, I would, providing he/ she takes into consideration both his / her experience and the information from the AI.

Would you share your personal data to help develop an AI for medical reasons?

Very probably IF I felt the medical team were trustworthy.

Would you trust a doctor using AI to make your medical diagnosis?

yes , I think if medical professionals were using this AI it would be very developed and probably make sensible decisions not effected by biases so I would trust it especially if a doctor was also present to over see decisions.

Would you share your personal data to help develop an AI for medical reasons?

I already share my data with every ‘free’ app (from Instagram to Gmail) I sign up to so why not? I’d like some assurance that the data would be anonymised though.

Would you trust a doctor using AI to make your medical diagnosis?

Yes - as part of a repertoire of resources

Would you trust a doctor using AI to make your medical diagnosis?

Yes because an AI is less likely to make mistakes compared to a doctor and may recognise markers earlier than a doctor can.

Would you share your personal data to help develop an AI for medical reasons?

I would share my personal data to help develop an AI for medical reasons. The first reason is that I would love to be a part of something that would make healthcare better for not only myself but others. The second reason is after seeing ted talks and reading articles on AI being more likely to have racial and gender bias I would like to be a part of the development process so in a small way I can reduce the amount of bias it may have.

Would you trust a doctor using AI to make your medical diagnosis?

Yes. Computer based analysis should be able to make objective outcomes to a higher level of accuracy than a human that is subject to errors, subjective opinions etc.

ARTIST

Bethany Hadfield

Bethany Hadfield is a London-based artist in her first year of the MA Painting program at the Royal College of Art. Beth is primarily a painter; she employs video and digital software in an experimental way to create motifs and environments in 3D spaces which are then translated onto canvas. Rendering these digital spaces in paint creates a sense of vitality. Beth is interested in the online body, where the physical body is compressed and uploaded, and we are represented through external features. She uses painting to explore the many interfaces of our digital selves on a macro- and microscale.

x

SCIENTIST

Jonny Jackson

Jonny Jackson is in the second year of his PhD at the CDT. The aim of his PhD is to identify signs of heart disease in real time using 2D images captured during surgery. With a background in Aeronautical Engineering, he brings forward his experience of using computing to model fluid dynamics in order to understand blood flow in the coronary arteries. Throughout his studies he has been involved with several public engagement projects that continue to remind him of the importance of sharing science with diverse communities to influence and inspire them.