MAM
Deepfakes target women in 93 per cent of cases, report finds
Pi-labs study shows 900 per cent rise in female-focused synthetic media; India sees 60 per cent jump in cybercrime complaints.
MUMBAI: Deepfakes aren’t just fooling cameras, they’re hitting women hardest, turning pixels into a new kind of weapon. A new report from creator intelligence platform Pi-labs has revealed that nearly 93 per cent of deepfake victims are women, with deepfake content targeting females surging 900% in recent years. The findings paint synthetic media as a fast-escalating digital threat with a stark gendered impact.
In India, cybercrime complaints involving women rose from about 50,000 in 2024 to nearly 80,000 by 2026, an increase of roughly 60 per cent in just two years. Almost 98 per cent of deepfake pornography is aimed at women, often powered by face-swapping apps and bot networks that disproportionately target females, including school-age girls. Victims typically fall in the 18–30 age group, with Bengaluru reporting a growing share of cases.
Globally, 62 per cent of deepfake abuse cases involving women go unreported due to stigma, in India, over one-third of women facing online harassment take no action, and many reduce their digital presence after abuse. Close to 33 per cent of women remain unaware of protective laws.
City-level trends show Bengaluru leading with nearly 30 per cent of complaints, followed by Hyderabad (14 per cent), Mumbai (13 per cent), Chennai and Kolkata (5 per cent each), and Delhi (3 per cent).
Pi-labs, CEO and founder Anukush Tiwari said, “AI is one of the most powerful technologies of our time, but like every powerful tool, it reflects the intent of those who use it. We are witnessing a growing trust deficit in digital spaces, where identity can be manipulated within minutes and reputations can be damaged overnight.”
Image morphing and deepfake videos remain the most common forms of misuse. The report also notes a new trend: fully AI-generated female personas (not based on real individuals) gaining high engagement on social platforms, raising questions about digital credibility.
Detection remains challenging due to widespread generative tools and rogue creators. Industry estimates suggest over 5,000 face-swap tools and more than 1,000 voice-cloning applications are accessible online.
pi-labs offers pi-authentify, an AI-driven detection system that scans media for generative markers and provides authenticity scores, as well as Namokavach, a verification portal delivering confidential assessments within two working days. The Payal gaming case was resolved using pi-authentify’s forensic analysis.
The report urges minimising digital footprints and adopting detection tools to limit replication risks. It frames the gendered impact of synthetic media as an urgent digital safety issue requiring coordinated action from individuals, platforms and technology providers.
In a world where faces can be borrowed in seconds, the real crime isn’t just creation, it’s the silence that follows, and women are paying the heaviest price.
MAM
Three senior OpenAI infrastructure executives join Meta
Key members of Stargate project move to rival amid aggressive AI spending race.
MUMBAI: Three key architects of OpenAI’s ambitious data centre plans have switched sides and joined Meta Platforms, according to people familiar with the matter. Peter Hoeschele, who played a central role in OpenAI’s high-profile Stargate initiative, is among the new hires. He is joined by Shamez Hemani, who focused on computing strategy and business development, and Anuj Saharan, another leader in the computing organisation. The Information first reported their departure from OpenAI on Thursday.
The moves come as Meta CEO Mark Zuckerberg has pledged to spend aggressively on AI infrastructure. The company is projecting capital expenditure of up to $135 billion this year alone, with hundreds of billions more expected before the end of the decade to support its Meta Superintelligence Labs and new models such as Muse Spark.
OpenAI, which is pushing ahead with massive data centre expansion, had described its early lead in securing computing power as a competitive advantage. Stargate, originally announced last year as a $500 billion venture involving OpenAI, Oracle, and SoftBank, has since become an umbrella term for the company’s broader data centre ambitions. However, the project has seen recent adjustments, including a pause on its UK plans and the decision not to expand the Abilene, Texas site.
A spokesperson for Meta declined to comment, while Hoeschele, Hemani, and Saharan also declined to comment. OpenAI said it was grateful for the contributions of the three employees and remains focused on hiring talent for its infrastructure plans. The company recently brought in former Intel executive Sachin Katti to lead its industrial compute efforts.
In the high-stakes race to build the future of artificial intelligence, talent is proving to be as valuable as computing power itself. Meta’s latest hires suggest the competition for top infrastructure minds is intensifying, even as OpenAI continues to scale its own ambitious projects. The move highlights how quickly the AI talent war is heating up across Silicon Valley.
