NEW DELHI: Google is ramping up security measures and research to combat online scams in India with an initiative called DigiKavach, the company announced on Thursday as part of a host of new collaborations and commitments, including the start of manufacturing of its flagship Pixel phones in the country.
The announcement on the security measures comes at a time of growing unease over a spurt in cyber crime, especially those involving financial frauds, such as instances of bank accounts being compromised and people being extorted by predatory loan apps.
Among the measures is a ramped up threat intelligence framework, which includes collaboration with fintech companies for on-ground feedback, and “deep research” into how scammers are evolving, which the company plans to share more widely to detail new scam patterns.
“We feel very confident that when we know a threat we can we can solve it. But what we have noticed is these bad actors are shape shifting,” said Saikat Mitra, vice president and the head of Google’s trust and safety for the Asia Pacific region, in an interview with HT on the eve of the company’s Google for India event on Thursday.
Mitra described the initiative as a “holistic hyperlocal, India focused threat intel capability”, which will take feedback from financial institutions, open source investigation (OSINT) techniques that study social media reports, and inputs from law enforcement and government agencies.
At the event on Thursday, the company announced it was partnering with FACE, or Fintech Association for Consumer Empowerment, to designate it a “priority flagger” — a collaboration that Mitra said will help the company get market intelligence.
Cyber scams in India have played out using methods that are both conventional – such as phishing, wherein people are fooled into clicking malicious links – but also novel, such as Google search listing hijacking in which fake customer care numbers are used to bait people into calling up scam call centres.
A more worrying vector has been the use of micro-loan phone applications often floated by fly-by-night companies. These have been used for extortion after the applications would siphon off their target’s personal data and images.
Mitra said Google launched a crackdown on such applications and made changes based on the experiences in India.
“In the latest change we did, which was a policy we implemented in May, I think, if you want to list a loan app, you do not get access to photos, you do not get access to contacts. But even then, bad actors tried to evolve and began listing apps as loan helpers or calculators,” he added, illustrating the cat-and-mouse chase that both tech companies and law enforcement have encountered in this area.
The company also announced that it will leverage its artificial intelligence models, especially those around generative text, to help Indians access information and business opportunity better.
Asked about challenges such as AI bias and hallucination – generative AI models like ChatGPT have often been accused of conjuring up fiction as fact — the Google executive acknowledged these as areas of concern and said work was evolving.
“There is a tremendous amount of work and resource we are investing in ensuring our generative AI products don’t dish out content that is unsafe, so that they do not say things that are not right. Generative AI learns from the universe of content that exists today and that has bias. And to give you an example, an article once said ‘Bard is great but boring’ but we kind of took that with a bit of pride — we wanted it to be a bit boring rather than edgy and positional,” he said, adding that Google was focused on retaining what it calls “neutral perspectives” and reference points to indicate the source of information to users.
The executive also spoke on measures for election safety in the post-generative AI boom. “To my knowledge, we’re the first company to launch a policy that says any political ad using synthetic media needs to prominently identify as synthetic media. We have gone into details and we lay down specifics of how prominently these must be displayed to users,” he said.
India and the United States are set to hold crucial elections next year and experts have in recent years flagged concerns over the proliferation of deepfake visuals and audio. Google has in the past said it was working on classifier systems to help flag any content that is AI-generated.