Imaging societies collaborate on new ‘ethics of AI’ guidelines

Representatives from a handful of major global imaging societies are collaborating on a “living document” that will outline a clearer set of ethics for the use of AI in radiology.

The American College of Radiology first published the consensus draft document—co-authored by the ACR, American Association of Physicists in Medicine, Canadian Association of Radiologists, European Society of Radiology, Radiological Society of North America, Society for Imaging Informatics in Medicine and European Society of Medical Imaging Informatics—in late February.

“Importantly, the group included trainees, patients and other stakeholders such as an ethicist from MIT,” Geraldine McGinty, MD, MBA, said in a March 1 blog post on the ACR’s website. McGinty is the chair of the ACR Board of Chancellors.

“But despite the wide-ranging backgrounds and expert input that created this draft, the writing group and our societies’ leaders are very clear that this is just that: a draft,” she wrote. “That’s why, having kept the writing group fairly small to facilitate an accelerated turnaround, it’s now critical to get extensive comments from the broader imaging and healthcare ecosystem.”

The ACR is encouraging experts to contribute their two cents here any time before April 15.

The consensus document, which ran 38 pages long as of March 5, is broken into three main sections detailing the ethics of data, ethics of algorithms and trained models and ethics of practice. In-depth, the 18 authors address data sharing, privacy, automation and algorithm biases, patient preferences, workforce disruption and the idea of machines making medical decisions.

“AI has noticeably altered our perception of radiology data—their value, how to use them and how they may be misused,” the authors wrote in the document. “Rather than simply understanding AI, radiologists have a moral duty both to understand their data and to use the data they collect to improve the common good, extract more information about patients and their diseases and improve the practice of radiology.”

The authors said it’s important radiologists start to develop these codes of ethics and practice for AI early, since “the radiology community is learning about ethical AI while simultaneously trying to invent and use it.” It’s also why McGinty predicts the code of ethics will remain a living document.

“We...anticipate that, given the pace of change in this sphere, this document will be a living one,” she wrote. “There is much work to do to ensure that we meet the goal that we’ve set for our ACR Data Science Institute: to promote safe, appropriate data science solutions that improve the care we provide to our patients.”

The ACR’s document comes just a week after the Royal Australian and New Zealand College of Radiologists, or RANZCR, published a draft of its own guidelines for the ethical use of AI and machine learning in healthcare. That statement, titled “Ethical Principles for AI in Medicine,” can be found in full here, and it’s open to feedback until the consensus period closes April 26.