Generative AI: A bilateral online safety focus for Australia and the UK

After delays due to generative AI developments, the industry code applying to search engine services (Industry Code) came into force on 12 March 2024. This follows the recent announcement of the Australian and UK governments joining forces in a Memorandum of Understanding (MoU) to strengthen bilateral cooperation to tackle online harms.

Both the Industry Code and the MoU share a focus on rapidly evolving technologies including generative AI.

You can see our previous articles relating to the online safety state of play here.

What is required under the Industry Code?

The Industry Code requires providers of internet search engine services to undertake certain compliance measures to prevent and mitigate online harm.

Some of the compliance measures include:

• regular reviews, testing and adjustments of machine learning algorithms and/or models to reduce the risk of class 1A and 1B material from being accessible to end-users (class 1A material comprises child sexual exploitation, pro-terror, and extreme crime and violence material, and class 1B material comprises crime, violence and drug-related material)
• delisting or blocking search results, links or cached copies to certain material
• deterrent messaging outlining the risk and criminality of seeking certain search results
• notifying eSafety of any significant changes to the functionality of the service that are likely to have a material effect on the risk of access or exposure to, or distribution of, class 1A or class 1B materials
• age-appropriate safety settings to limit exposure to explicit and/or graphic material
• notifying law enforcement if certain child-exploitation material is identified on the service.

What are the consequences for non-compliance?

Failure to comply with an industry code carries a financial penalty of AUD$156,500.

What are the focus areas of the MoU?

The Australian and UK governments’ MoU lists key focus areas which include age assurance, encryption, misinformation, and the impact of new, emerging and rapidly evolving technologies such as generative AI.

The MoU outlines an intended approach which will involve:

• collaboration and cooperation between law enforcement agencies and regulators
• coordinated engagement with global tech companies
• sharing best practice to counter misinformation, disinformation and foreign interference
• supporting efficient routes to market for safety technology solutions and innovations
• amplifying existing specifications and standards to ensure safety, privacy and security are built in by design for new and emerging tech (including AI).

What is the impact of the MoU?

While the MoU does not contain any binding obligations, it outlines the intended bilateral approach of the Australian and UK governments in tackling online harm. We expect that the MoU will inform the eSafety Commissioner’s approach to engagement with industry and regulatory enforcement in the future.

What next?

Online service providers should familiarise themselves with the industry codes and stay up to date with developments in the online safety space.

As flagged in our previous article, we expect the eSafety Commissioner to publish the enforceable industry standards for ‘relevant electronic services’ and ‘designated internet services’ services soon, following the consultation period which closed on 21 December 2023.

In addition, we expect amendments to the Online Safety (Basic Online Safety Expectations) Determination 2022 focusing on generative AI and user empowerment to be published soon, following the end of the consultation period on 16 February 2024.