Meta deployed regional AI models over the past year to identify sensitive material and restrict it from recommendation feeds in India. While posts that comply with overall platform rules can stay online, they are blocked from being promoted to unfollowed audiences. User complaints about suggested content have reportedly dropped by half, though Meta did not share the underlying data. These details follow an October advisory from India's IT ministry directing Meta to evaluate its recommendation systems.
Key points
Compliant posts containing violence or sexually suggestive themes can remain hosted but w…: Compliant posts containing violence or sexually suggestive themes can remain hosted but will lose algorithmic recommendation status.
India's IT ministry directed Meta in October to review algorithmic ranking for misleading…: India's IT ministry directed Meta in October to review algorithmic ranking for misleading and violent content.
Meta did not state whether it altered its recommendation algorithms regarding specific re…: Meta did not state whether it altered its recommendation algorithms regarding specific real-world events in India.
Sources checked by our analysis pipeline: 1
Meta, which owns Instagram, said the platform now recommends fewer sexually suggestive posts and depictions of violence. (Representative image)
NEW DELHI: Meta has tightened Instagram's recommendations in India, disclosing the changes a month after the govt asked it to review how Instagram and Facebook suggest posts. Meta said Instagram now recommends fewer sexually suggestive posts and depictions of violence.The changes apply to suggested posts from accounts users do not follow, including in Reels, and the main feed. Posts complying with Instagram's rules can remain on the platform even if they are excluded from recommendations.Meta said it had strengthened its default safeguards over the past year based on user feedback. Instagram uses AI models to assess posts before recommending them and has fine-tuned its systems for India to identify more sensitive material."We've seen people report recommended content about half as often as they were a year ago," Meta said in a blog post. The company did not disclose the underlying figures.In its Oct 10 advisory, the IT ministry asked Meta to review how it ranks and recommends posts, citing concerns that…
Threat actors leveraged the digital experience platform to distribute a combined malware cocktail to users browsing a wide array of affected domains. The multi-pronged payloads are engineered to conduct digital asset theft while simultaneously deploying dedicated mobile exploits against visitors using Apple smartphones. This compromise significantly heightens risks for web users navigating normally trusted online destinations.
A routine internal software change disrupted multiple consumer and enterprise platforms on Tuesday, interrupting services ranging from media streaming to digital payments. The incident compounded a turbulent week for digital infrastructure as separate corporate platforms struggled with technical stability. Companies involved have consistently attributed these widespread interruptions to operational errors rather than malicious cyberattacks.
The findings mark a noticeable reversal for Apple after recent iterations of the iPad and HomePod showed improvements in product disassembly. Technicians suffered component damage while opening the unit due to aggressive adhesive and risky internal thermal bonds. As a result, the smart hub poses severe servicing challenges that contrast sharply with the company's recent sustainability and repairability progress.