Meta’s Instagram faces fresh scrutiny over “algorithmic cravings,” as a safety advocate compares it to a “digital cigarette”

Instagram is again drawing analogies from the digital-safety world, with a prominent advocate arguing that the platform’s recommendation systems can turn ordinary browsing into something closer to compulsion.

In comments cited by Fortune and republished by Yahoo Finance, Larry Magid compared Instagram to a “digital cigarette” alongside a “box of chocolates,” suggesting the experience can encourage users to keep consuming rather than step away. Magid’s point, as described in the piece, is that the “algorithms change the whole nature of it,” reshaping how people encounter content over time.

The metaphor is unusual but not random: it draws attention to the difference between a feed that simply shows what a user chooses and a feed that continually selects what it predicts the user will click next. Critics often focus on that selection layer, arguing it can drive rapid feedback loops that keep users engaged even when the content is emotionally charged or habit-forming.

Magid’s remarks also suggest a practical problem, not just a moral one. If the same app is used for entertainment and social connection, an algorithmic layer that optimizes for attention can make the product behave differently than users expect, the way a familiar ritual can take on a different character once a system nudges behavior.

How to “fix that,” as the headline frames it, is where the debate usually turns technical and policy-based. Recommended-content systems rely on models trained on indicates like watch time, repeat views, shares, and reshares, then reshuffle the ranking to maximize likely engagement. The question for platforms is what objective should dominate those rankings, and what constraints, friction, or redesigns should be introduced when engagement can conflict with safety goals.

Meta has long treated Instagram as part of a broader family that includes Facebook and WhatsApp, and it has also expanded public discussion around safety, youth well-being, and AI systems that personalize feeds. Still, this particular report does not lay out new, specific changes by Meta, and it does not describe any concrete product redesign being announced in response to Magid’s comments.

For the business side, the tension is familiar. Instagram’s recommendation technology is central to the platform’s growth and ad targeting, and it is also the main lever that determines how often users encounter new content. Any effort to reduce compulsive loops can come with tradeoffs, including changes to time-spent patterns, reach, and ad inventory assumptions.

What remains unclear from the cited remarks is the scope and accountability of any remedy. The report attributes the concerns to a safety advocate, but it does not provide new data, specify measurable outcomes, or confirm whether Meta is planning algorithmic changes tied to compulsive consumption risk. That means readers are left with a critique and a question, rather than a documented policy response or timetable.