Monash expert: Digital Duty of Care, opt-in v. opt-out and emerging AI harms
Monash University
A Monash University expert is available to comment on the proposed Digital Duty of Care legislation expected to be released later this week, and the need to consider emerging AI harms to broaden the scope of online protections.
Professor Rashina Hoda, Department of Software Systems and Cybersecurity, Faculty of Information Technology
Professor Hoda was part of the recent Design Dialogue on the Digital Duty of Care organised by the Tech Policy Design Institute at the Parliament. She has also written a policy brief on future-proofing the Duty and authored an op-ed for the AFR on the ‘AI gender gap’.
Read more of Professor Hoda’s commentary.
Contact via: +61 450 501 248 or [email protected]
- Human and socio-technical aspects of software engineering
- Responsible AI
- Digital health
- Encouraging entry and retention of women into STEM fields
The following can be attributed to Professor Hoda:
“The Australian Government is expected to release draft legislation creating a digital duty of care for social media users to strengthen online safety and protect Australians from harm.
“While an opt-out option for recommender algorithms would put more control in the hands of Australian internet users, it may not always be easily accessible or implementable.
“For example, users may not even know they are in an echo chamber and therefore may not be able to help themselves. This could be even more challenging for vulnerable groups, including young people, the elderly and users with mental health conditions.
“Although it comes with its own challenges, such as the loss of personalised content, an opt-in option would provide a stronger control setting for users.
“Improving digital safety in social media contexts such as Instagram and Facebook is a good next step. But if the new amendments do not already include safeguards against harms increasingly being caused by modern generative AI systems such as ChatGPT, Gemini and Claude, these should also be seriously considered.
“These include AI-generated deepfakes, which overwhelmingly target girls and women, as well as AI sycophancy and the emerging phenomenon of ‘AI psychosis’ associated with conversational agents or chatbots. These can also be particularly harmful for our vulnerable populations.
“Generative AI also allows users to create potentially harmful content, rather than harm arising only from social media algorithms curating and feeding users harmful content pre-existing online. We are also starting to see problematic behaviour from AI agents when they do not follow the rules they have been given.
“These emerging and rapidly developing AI harms also need to be considered as part of the broader conversation about digital and online safety.”
For any other topics on which you may be seeking expert comment, contact the Monash University Media Unit on +61 3 9903 4840 or [email protected]