Adam Mosseri, the head of Instagram, recently addressed concerns about the platform’s recommendation system during an Australian press briefing. He revealed that users who disable Instagram’s algorithmic feed experience a significant decline in engagement—up to 50%—and a sharp drop in satisfaction scores. Mosseri explained that switching to a chronological feed often results in users being overwhelmed by content from brands, companies, and publishers, rather than posts from friends and personal connections. This leads to users missing more content and spending less time on the app.

This statement comes as Australia considers a Digital Duty of Care bill that would require platforms to prompt users to turn off ranking algorithms if they choose. The legislation is motivated by concerns that increased engagement and time spent on platforms may contribute to harm, a point Mosseri’s warning indirectly highlights.

Mosseri’s argument focuses on engagement metrics rather than user wellbeing, which is the core concern of the proposed legislation. The issue is complicated by Instagram’s own design choices, as the platform’s algorithm has historically encouraged high-volume posting by professional accounts. Simply removing the ranking system does not address this behavior.

Meta has previously promoted its AI-driven recommendations as a way to increase user time on Instagram, presenting this as a positive outcome. Mosseri’s recent comments frame reduced time on the platform as a negative, underscoring the ongoing debate about whether increased engagement benefits users.

Instagram also plans to make algorithm customization more accessible, aligning with Australia’s proposal to prompt users about feed ranking controls. The disagreement lies in how and when users are offered this choice, rather than whether they should have it.

Instagram reintroduced a chronological feed option in 2022 after political pressure in the United States, but it was not a persistent setting and had to be selected repeatedly. The Australian bill aims to ensure that such options are actively presented to users, as rights that are not surfaced tend not to be exercised.

Notably absent from Mosseri’s remarks were any details on wellbeing metrics, such as user regret or emotional impact, which are crucial to assessing whether changes to the feed improve or harm users. The 50% drop in engagement indicates a change in user behavior but does not clarify the effect on user welfare.

Regarding Instagram’s potential addictiveness, Mosseri distinguished between colloquial and clinical definitions, reflecting the company’s cautious stance amid ongoing litigation.

Europe’s experience may offer insight, as the Digital Services Act requires large platforms to provide a non-profiling feed option. Adoption there has been limited, which some interpret as user disinterest. However, Mosseri’s engagement figures suggest that the low uptake may be due to how hidden the option is rather than lack of demand.

The debate over Instagram’s feed algorithm highlights broader questions about user control, platform responsibility, and the impact of engagement-driven design on wellbeing.