'Digital platforms must implement prevention by design to prevent CSEAM'

Experts urge internet platforms to adopt prevention-by-design and stronger AI systems to curb child sexual exploitation material and improve online child safety

Internet and social media intermediaries must implement a prevention-by-design principle to prevent the proliferation of Child Sexual Exploitation and Abuse Material (CSEAM) on their platforms, according to experts.

 

While most platforms treat the detection of CSEAM as a content moderation challenge, the most significant challenge in catching and removing such content is contextual interpretation and the lack of known databases against which it can be matched, the experts said.

 

"AI systems can match known CSAM against hash databases with reasonable accuracy, but they consistently fail at understanding intent and context in novel material. An image that appears benign in isolation can be part of a grooming sequence when viewed across a conversation thread," said Nilesh Jahagirdar, co-founder and vice-president, marketing and solutions, at [x]cube LABS.

 

Malicious actors involved in the spread of CSEAM constantly evolve their tactics faster than detection models can adapt, said Parminder Singh, co-founder of RedScope AI, a decision intelligence platform.

 

"AI is exceptionally good at identifying known patterns, but child exploitation networks deliberately operate in the grey areas, using coded language, edited visuals, screenshots, memes, private groups, ephemeral content and multilingual communication to evade detection," he said.

 

A second problem that most platforms face is the lack of investment in multilingual AI capable of understanding India's linguistic diversity, and in continuously evolving their detection systems as abuse tactics change, the experts said.

 

"Most detection models are trained predominantly on English-language data. Regional languages, coded slang and local dialects used by bad actors fall almost entirely outside detection parameters," Jahagirdar said.

 

Digital intermediaries should also invest in reporting offenders who circulate CSEAM to law enforcement agencies, rather than just disabling their accounts and removing the content from their platforms, Bhuwan Ribhu, founder of Just Rights for Children, said.

 

"Online child sexual exploitation is a borderless crime and demands a borderless response. That requires international cooperation in sharing intelligence, digital evidence and financial information. Until governments, technology companies and digital intermediaries accept this shared responsibility, no child is safe online," Ribhu said.

 

Earlier this month, Meta's video- and photo-sharing platform Instagram drew the ire of the Ministry of Electronics and Information Technology after reports surfaced that it was allowing advertisements promoting CSEAM.

 

The IT ministry summoned senior executives from the company and asked them to explain how such content was being allowed on the platform.

 

Subsequently, in a blog post published on Tuesday, Meta said it had removed 160,000 accounts in India for hosting and sharing "suspicious off-platform links in coordination with other signals indicating child exploitative activity".

 

Overall, Meta, which has WhatsApp, Facebook, Instagram and Threads under its umbrella, said it automatically removed more than 4 million suspicious accounts from Facebook and Instagram, in addition to 36 million pieces of content containing child exploitation, according to the blog post. 

Expert view

• Challenge in catching and removing abusive content is contextual interpretation
• Malicious actors constantly evolve their tactics faster than detection models
• Lack of investment in multilingual AI capable of understanding India’s linguistic diversity