How to Tell If Your Partner Used AI to Wishing You Happy Valentine’s Day

Valentine’s Day is already here, and for many it will be a holiday marked by rage generative artificial intelligence. Millions of lovers around the world will surely use this technology to create something unforgettable for their loved ones. From greeting images to poems and other greetings, AI has proven more than effective in creating them.

Fine? This is not true? Frankly, we cannot make value judgments on such issues. Artificial intelligence is more accessible to everyone today than ever before. So it makes sense that the public uses this to dedicate a message to the person they share their life with, or someone they would like to do so with. In some cases, technology can help express in words or pictures what many find difficult to express otherwise.

However, if for some reason you want to know if the Valentine’s Day greeting you received was generated by artificial intelligence, There are different ways to find out.. Below we will tell you some basic parameters to consider, as well as tools that can help you with this task.

Valentine’s Day and greeting cards created using artificial intelligence

An original card is always a good gift for a holiday like Valentine’s Day. And a huge number of image generators based on artificial intelligence make it very easy to create them in a matter of seconds. DALL-E 3from OpenAI, Middle of the road or Microsoft designerare one of the most popular tools for this task.

A simple text description will be enough for any of these platforms to create an impressive image. But how do you know if the one your partner gave you for Valentine’s Day was created using artificial intelligence? The easiest way to find out is look for the watermark in the lower corners of the map.

Both Designer and DALL-E 3 have a small logo that allows you to instantly identify if they were created using these applications. Although it is also true that They are easy to cut or remove. using programs such as Photoshop, GIMP or Photoscape.

You can also look for flaws in the items that are depicted on the valentine card. Odd shapes, misspellings, and even poorly implemented body parts are often direct signs of AI being used. Midjourney, for example, has become one of the best at creating realistic images. However, he has difficulty handling people’s hands. So, for example, if you see a picture of folded hands, most likely some of them have 6 fingers. It sounds stupid, but it’s a good sign to keep in mind.

Poems and messages about love

If your partner gave you a poem or other written messages for Valentine’s Day, you have more tools to try to recognize if they were created by AI. Unlike images, which require visual inspection, there are several online applications that can detect if text was created using ChatGPT, Gemini or Copilot; or even if it is a hybrid of human creation and artificial intelligence.

But be careful, because although some of these utilities are over 90% effective, they can fail, for better or for worse. This means they can mark text as being written by a human when it actually came from an AI, and vice versa.

Chatbots like Gemini, Copilot or ChatGPT have become very good at creating artistic or very sentimental content. You can even ask him to make personalized creations for Valentine’s Day, such as: “Write a love poem for Mary from Luis”; or repeating the work of famous authors, for example, “Write a love poem in the style of Mario Benedetti.” This makes it increasingly difficult to determine its authenticity.. Don’t forget that OpenAI launched its own app to check if text was generated by AI, but scrapped it due to inefficiency.

Utilities for detecting traces of artificial intelligence in Valentine’s Day greetings

If you think your partner wished you a Valentine’s Day poem written by artificial intelligence, you can use these tools.

• GPTZero– An open source solution that validates user-entered text and matches it against various language models. It can detect if it comes from ChatGPT, other applications with GPT-4 and LLaMA 2.
• undetectable AI: An alternative that analyzes text and compares it to other similar apps looking for possible AI clues. It includes the ability to “humanize” text, which removes clues about whether it was created by artificial intelligence.
• Piggy Banks: Like the above, it looks at entries provided by users and determines whether they were created using ChatGPT, GPT-4 or Gemini.

As we already said, None of these options are reliable.. But at least they can help you know whether the Valentine’s Day greeting you received from your loved one was an original creation or not.