Close up of a person's eye

Using gliff.ai to Realise the Potential of OCT Data

The aim of the OCT data project is to create a ground-truth annotated dataset that would be utilised for the future development of fully automated AI tool capable of segmenting and quantifying different features in retina images and scans…

‘Coded Bias’ brings bias in AI into mainstream

North-East company, gliff.ai, founded in 2020, develops tools to support the creation of Artificial Intelligence (AI) – particularly for healthcare, life science and manufacturing.

On 5th April, Netflix UK released the documentary film Coded Bias, originally premiered in 2020 at the Sundance Film Festival. giff.ai feels that this film is helpfully bringing issues regarding bias in data science into mainstream media.

Coded Bias primarily focuses upon how institutional racism and sexism is embedded into datasets and subsequent algorithms, based upon the research of Joy Buolamwini from MIT. The film successfully demonstrates why unregulated, ill-thought-out and opaque approaches to developing AI can easily lead to flawed models that may not only discriminate against certain groups of society but also fail drastically to meet the applications’ original objectives.

However, bias in AI is far from limited to demographic-based data, facial recognition technology and civil-rights issues. Bias can appear in machine learning in many ways and result in myriad consequences, including effects that may be incredibly difficult to detect or understand. Just last month, a research paper found that bias was among the reasons why 2,212 machine learning models for detecting and diagnosing Covid-19 were not fit for clinical practice. 

Furthermore, race and gender equality in healthcare remains a challenge in medical technology, as recently highlighted by The Economist which points to the example of pulse oximeters to show how non-white people and women may be discriminated agaist via the use of medical devices. The article states: “[…] the wider point is that medical technology should be designed from the outset to be free from such bias—and, unfortunately, it isn’t.”

This is why the mission of gliff.ai is to tackle issues such as bias in AI head-on.
“At gliff.ai, we firmly believe that AI should be fair, accountable, transparent and ethical. To assist our clients who are developing AI models, gliff.ai offers innovative, open-source software tools that are built upon these values,“ says Chas Nelson, gliff.ai’s Chief Technology Officer.

“AI has incredible potential to improve our world, for example by helping doctors to effectively diagnose patients early on. But first, we must build trustworthy AI. Many people are unaware of the challenges involved in developing fair and ethical AI models. The documentary Coded Bias, however, provides an insightful introduction into this topic and is helpfully bringing issues regarding bias in data science into mainstream media. We hope that this will educate more people about the importance of ethics in AI.”