How to tackle threats of AI dress remover bot, deepfake video tools

A recent BBC report highlighted that the risk of harassment through social media is increasing at an alarming rate in Bangladesh. The report underscores that women and minors are predominantly targeted in these cases.

The use of artificial intelligence, including AI dress remover bot, deepfake video tools, for targeted harassment has led to a situation where women are hesitant to share photos on social media.

In a report, bdnews24.com pointed out bitter experiences suffered by Bangladeshi girls. The report said a resident of Chattogarm, Sultana Razia [not her real name], recently found herself thrust into a scary ordeal.

“Around two weeks ago, she received an unsettling message from an unfamiliar sender on her Instagram account. She was shocked to see a distressing image that had manipulated a normal photo of herself dressed in traditional attire into something explicit,” the report said.

The sudden incident left her deeply unsettled and unsure about how to cope.

The report said another victim, a female Bangladeshi TikTok star and YouTuber, shared a similar tale of harassment.

A nine-second explicit video featuring her started circulating on different platforms, prominently on Facebook, where it gained significant traction.

On Aug 5, the TikTok star took to Facebook and released a seven-minute video statement, clarifying that the video had been created using AI technology. She also said everyone struggles to cope with such situations in this social context.

The victim also got threatening emails for ransoms, which said, “From now on, when I edit videos, there will be no tattoos in them.”

"They demanded $3,000 and warned, 'If you don't pay, my editing will get even better.'"

"I provided the email copies to the cybercrime authorities. They are tracking the IP address of the threat.

“I've found out it's being operated from another country, but someone in our country is involved," she said.

HOW IS IT DONE?

The Telegram Dress Remover Bot utilises sophisticated AI algorithms to remove clothing from images containing individuals digitally.

While this tool holds intriguing potential for image editing, it's vital to understand the ethical and legal issues it brings forth.

Speaking to bdnews24.com, Sumon Ahmed Sabir, an expert in information technology, explained that bots specifically pertain to software integrity and can be hosted on diverse servers or computers.

This type of bot has the capability to execute tasks by establishing a connection from a remote location.

For instance, he said, chatbots are widely used in social marketing, and when there is a high volume of daily complaints in customer care, chatbots can handle routine queries.

 “They ask users about their needs and transfer the conversation to a human when necessary. This has significantly reduced the workforce needed in call centres.”

Previously, a call centre might have required 1,000 employees, but now, with the help of bots, the same work can be done by just 20 people, he said.

But there are also harmful variations of bots like malware bots, said Sumon, the chief technology officer of international internet gateway operator Fibre@Home.

 “Imagine you download software or visit a website, and a bot ends up on your device.

"These malicious bots can record what you type, including private details like usernames and passwords. They mimic human actions but with harmful intentions."

Some bots have the capability to transform photos quickly, and they can even convert photos into videos.

"Advanced AI graphics technology can effortlessly modify both photos and videos. Like blending faces in photos – this capability also expands to videos," Sumon said.

"Previously, editing images using Photoshop was a time-consuming process. If you weren't proficient in Photoshop, turning a regular photo into a fake nude one was complicated. Nowadays, by adding a face to a video, that face is automatically integrated throughout the entire video. You might not notice it's fake if you're not careful."

HOW TO TAKE ACTION, REPORT

StopNCII.org is a free tool designed to support victims of Non-Consensual Intimate Image (NCII) abuse.

Victims are uploading actual and AI-generated images to the website stopncii.org to get them removed, and the website is also providing free removal of AI-generated images from all online platforms.

Meta Platforms, the owner of Facebook, advises taking the following steps if someone has shared or threatened to share intimate images.

• Contact local law enforcement if concerned about physical safety.
• Consider seeking professional support and guidance.
• Explain the situation to a trusted friend or family member for support while navigating the process.
• Take screenshots and print out pages of photos and threats before taking any steps to delete the images, as these may be needed as evidence for pursuing legal action.

WHAT POLICE SAY

Mohammad Saiful Alam Khan, director of Bangladesh government's e-Government Computer Incident Response Team, or BGD e-GOV CIRT project, said: "I'm aware of this, although I haven't conducted a detailed analysis yet. We've heard that these incidents are happening with the help of AI tools from other countries.”

"I heard it's happening on Telegram too. However, we must conduct a thorough investigation, as we are currently looking into the matter. We can't say anything for sure without analysing it."

Md Najmul Islam, additional deputy commissioner at the Cyber Crime Investigation Division of Dhaka Metropolitan Police, said, "We have been informed about this situation and are aware of it. However, as of now, we haven't received any complaints about individuals being harmed or victimised."

"This is a recent concern. I strongly recommend that those affected should promptly inform the police. The police will take the necessary action. Through analysis, our forensic lab can confirm if the content is real or fake."

When asked if the police contacted Telegram regarding this matter, he replied, "We'll respond if there's an incident. Otherwise, we won't initiate contact.

 “We lack information about the victims or those facing issues. Reporting a problem provides us with leads. The police will then pursue these leads."