Google Photos adds AI indicators

Estimated reading time: 3 minutes

For now, AI images look so realistic that it is difficult to distinguish them from real-life photos. Google has this in mind, although we have not yet seen whether it will work.

New adjectives are being introduced in Google Photos, making it easier to separate photos that have been altered with the help of its already popular features like Magic Editor and Magic Eraser. This move will serve as a tool to inform users when Artificial Intelligence has been used to manipulate an image – handy at this time considering generative image tools are now common and Google has even incorporated a host of AI editing features into the Photos app itself.

Hitherto, these AI editing tools were only mentioned in the metadata that was contained in the image file. Now it’s easier also for the users since the info label will appear directly in the Google Photos app besides the other image details such as photo name and location. This labelling will apply to both generative AI used in editing that produces new image content and other non-generative AI in such functions as “Best Take”.

Read more on NRTC

• OnePlus went ahead and built its own version of Google Magic Eraser
• Google Photos has introduce Automated video editing
• 5 Essential Tools for aspiring Content Creators
• Chimamanda writes US, UK leaders, slams them for congratulating president-elect, Bola Ahmed Tinubu
• Election Results: Obasanjo write Buhari, says danger looming ahead

Google provides this move based on its dedicated AI principles as the key reason. The company is going to research additional ways to address its goal of increasing the understanding of how AI operates in image editing.

Sadly, this is not always an ideal way of figuring out whether something was built with AI or not. This takes the metadata of an image which should indicate whether an image was AI-generated or AI manipulated and the thing is that the metadata of an image is changeable. It is a thing that is openly offered as a disclaimer in this application. Also, if it took visual signals or hints in an image, then it is just as basic as erasing or cutting out those signals or indications for use in eradicating any sign of AI in the image—cues that Google confessed were enough to successfully discern if an image was AI-manipulated or not. Nevertheless, it’s always preferable to have no communication at all.