Social media firms need time to develop anti-deepfake tools: industry executives
Social media firms need time to develop anti-deepfake tools: industry executives
Social media and internet intermediaries, including the likes of Google and Meta, are likely to take more time to develop tools to accurately detect and take down deep fake and synthetic content that is present on their platforms, senior executives at these companies told ET.
At present, all content which is either problematic or needs to be taken down in accordance with the laws of the country needs to either be flagged by the user or gain “a certain level of traction or virality” to be recognised by machine parameters or artificial intelligence-driven identifiers, these executives said.
“We rely on our users to flag such content to take them down. Our identifiers for any type of content, let alone deep fake or synthetically altered, can only work if such content achieves minimum traction on the platform. Otherwise, it is impossible for anyone, let alone us, to detect any content and weed it out or take it down,” a senior executive at a social media conglomerate said.
The possible delay in developing accurate tools to tackle deep fake and other synthetic content is likely to put these companies in the crosshairs of the ministry of electronics and information technology.
Earlier this month, union electronics and information technology minister Ashwini Vaishnaw and minister of state for electronics and information technology Rajeev Chandrasekhar met senior executives from social media and internet intermediaries on consecutive days to discuss the issue of deep fakes.
Both Vaishnaw and Chandrasekhar had, after their meeting with the companies, said that deep fakes were a threat to democracy and stable governance and that they must be tackled immediately.
On one hand, while Vaishnaw proposed bringing new legislation to deal with deep fakes, Chandrasekhar said that all companies must, to begin with, notify users at all stages of their interaction with the platform's interface, the perils of uploading deep fake and synthetic media.
“We understand that this is a new technology and a comprehensive mechanism to effectively deal with it may take some time. In the interim, however, companies must either change their policies suitably or upgrade the existing tools to fend off the problem likely to be caused by deep fake content,” a senior government official said.
Almost all social media and internet intermediaries follow the same template when it comes to dealing with problematic content such as deep fake or synthetically altered content, the executive said.
“There are multiple components to how synthetic media is tackled. The very obvious ones are looking for account behaviour, when was the account which shared the content created and others. We also work with fact-checkers for a deeper, more extensive check of such content,” another executive at an internet intermediary said.
The most effective way, social media executives, government officials and experts said, was to educate the user and “clearly define what constitutes manipulated content” and the consequences for violating policies.
“It’s important to strengthen user verification, educate users on recognising deep fakes, and provide reporting mechanisms,” Soumen Datta, partner at consultancy firm BDO India, said.
While the government advisories and meetings with social media and internet intermediaries underscore the need for a “swift response”, and “real-time monitoring”, companies will also need to make the users aware through comprehensive awareness campaigns to “foster a digitally literate society capable of discerning the nuances of deep fake risks,” said Sunil Sharma, India vice-president of sales at Sophos.
Source : The Economic Times
At present, all content which is either problematic or needs to be taken down in accordance with the laws of the country needs to either be flagged by the user or gain “a certain level of traction or virality” to be recognised by machine parameters or artificial intelligence-driven identifiers, these executives said.
Elevate Your Tech Prowess with High-Value Skill Courses
| Offering College | Course | Website |
|---|---|---|
| MIT | MIT Technology Leadership and Innovation | Visit |
| IIM Kozhikode | IIMK Advanced Data Science For Managers | Visit |
| IIM Lucknow | IIML Executive Programme in FinTech, Banking & Applied Risk Management | Visit |
“We rely on our users to flag such content to take them down. Our identifiers for any type of content, let alone deep fake or synthetically altered, can only work if such content achieves minimum traction on the platform. Otherwise, it is impossible for anyone, let alone us, to detect any content and weed it out or take it down,” a senior executive at a social media conglomerate said.
The possible delay in developing accurate tools to tackle deep fake and other synthetic content is likely to put these companies in the crosshairs of the ministry of electronics and information technology.
Earlier this month, union electronics and information technology minister Ashwini Vaishnaw and minister of state for electronics and information technology Rajeev Chandrasekhar met senior executives from social media and internet intermediaries on consecutive days to discuss the issue of deep fakes.
Both Vaishnaw and Chandrasekhar had, after their meeting with the companies, said that deep fakes were a threat to democracy and stable governance and that they must be tackled immediately.
On one hand, while Vaishnaw proposed bringing new legislation to deal with deep fakes, Chandrasekhar said that all companies must, to begin with, notify users at all stages of their interaction with the platform's interface, the perils of uploading deep fake and synthetic media.
“We understand that this is a new technology and a comprehensive mechanism to effectively deal with it may take some time. In the interim, however, companies must either change their policies suitably or upgrade the existing tools to fend off the problem likely to be caused by deep fake content,” a senior government official said.
Almost all social media and internet intermediaries follow the same template when it comes to dealing with problematic content such as deep fake or synthetically altered content, the executive said.
“There are multiple components to how synthetic media is tackled. The very obvious ones are looking for account behaviour, when was the account which shared the content created and others. We also work with fact-checkers for a deeper, more extensive check of such content,” another executive at an internet intermediary said.
The most effective way, social media executives, government officials and experts said, was to educate the user and “clearly define what constitutes manipulated content” and the consequences for violating policies.
“It’s important to strengthen user verification, educate users on recognising deep fakes, and provide reporting mechanisms,” Soumen Datta, partner at consultancy firm BDO India, said.
While the government advisories and meetings with social media and internet intermediaries underscore the need for a “swift response”, and “real-time monitoring”, companies will also need to make the users aware through comprehensive awareness campaigns to “foster a digitally literate society capable of discerning the nuances of deep fake risks,” said Sunil Sharma, India vice-president of sales at Sophos.
Source : The Economic Times