Is it safe to use bot AI?

Today, I received this message:

Mr. Shubham is asking whether it’s safe to use AI bots. This is a complex question from both an informational and ethical standpoint, so I’ll try to give you a detailed answer.

What are AI bots?

AI bots are programs that use artificial intelligence to generate responses to user questions. The most well-known example today is ChatGPT (https://openai.com/chatgpt/), and on our website, the most famous example is DeepNude Telegram Bots. ChatGPT generates a text-based response to your question using LLM (Large Language Models). Meanwhile, DeepNude Telegram Bots respond to a photo you send by generating an image (Pix-2-Pix GAN) of the model undressed or altering their clothes.

Is it safe to use AI bots?

In general, my answer would be “Yes”, it’s safe to use AI bots, provided you follow the rules and have the right expectations.

Why are people afraid of AI Undress bots?

Data Theft

When I use the internet, I fear that my personal data might be stolen. The fact is, your personal data can be stolen even from the most secure databases, like well-known banks (Capital One in 2019, Banco de Chile in 2018, JP Morgan Chase in 2014).

So, a Telegram bot, like any other system on the internet, can be hacked or otherwise cause a data leak.

How can this be prevented?

• Don’t use your personal mobile number for your Telegram account.
• Buy a prepaid number and use that Telegram account for bots.
• Enable settings that hide your number and other personal information on Telegram.

My photos being leaked from an AI bot’s database

There’s concern that the developer of a Telegram bot might store the photos users send in order to collect them or blackmail the owner.

In theory, this is possible, but in practice, developers don’t do this. Because it would be an enormous amount of data. On average, the AI bots in our list are used at least 10,000 times a day. That’s 10,000 photos per day, which is more than 20 gigabytes. This is a lot of data that requires expensive hard drive space. Developers create these apps to make money, not spend it. That’s why popular Telegram AI bots don’t store photos on servers; they use your device’s memory instead.

How can this be prevented?

• Upload photos without your face.
• Use well-known AI bots with strong brands that care about their reputation.
• Don’t upload photos that are easily modified by AI tools: clear, front-facing images, minimal clothing.
• Upload photos where you’re wearing loose clothing and posing in complex ways, making it harder for AI to recognize the body.

Ethical aspects of AI

Personally, I’m against using any technology for malicious purposes. I believe the improper use of AI bots and apps should be punished. However, I’m also in favor of openness and accessibility to these technologies so that anyone can test them. It’s enthusiasts like these that push forward advancements in cutting-edge technology, whether it’s robotics, video games, or AI.

Legality of Use

August and September of this year have been pivotal in the creation of the anti-Undressing Apps movement. There are now rulings or ongoing legal processes against both the owners of these bots and their users. Here’s what I know so far:

• A ruling from a Spanish court against a user who used the results to harass an acquaintance.
• The beginning of a legal process in San Francisco against 16 websites and the companies that run them.
• The start of a legal process in South Korea against the bots and websites.

Based on these facts, I do not recommend using AI bots in any of these locations with other people’s photos.

Final words

In summary, based on the thoughts mentioned above, I conclude that for now, AI applications should only be used with images for which you have permission, and solely for personal use. Any other cases could expose you to legal risks, at least in the locations mentioned.

But if you don’t want your photos to leak, don’t upload photos where the model’s face is visible. Crop the photo or use a mask before sending it to the bot.