It seems like Instagram is working on new notifications to clarify when AI has been used to make posts.The snapshot was taken by app researcher Alessandro Paluzzi, who often finds new Instagram features before they are officially acknowledged or launched, and it states, “The creator or Meta said that this content was created or edited with AI.” In this example, the warning says the image was “generated by Meta AI” and then explains what that means and how to recognize AI-created content in the future.
This finding follows recent pledges made to the White House by Meta and other key AI firms such as Google, Microsoft, and OpenAI to ensure the ethical advancement of AI. One of the pledges was to create a watermarking mechanism to alert viewers when a material is AI-generated, in addition to funding cybersecurity and discrimination studies.
Instagram’s labelling mechanism, if any, remains a mystery, as does the degree to which it will rely on users to reveal when artificial intelligence was employed in creating or editing a picture. The inclusion of “Meta said” in the notification, however, implies that the tech company will actively apply the notice in at least some circumstances rather than depending on the users’ honesty. Neither Meta nor a representative for the firm immediately responded to requests for comment from Engadget nor The Verge.
AI-generated disinformation is still in its early stages. Still, we got a taste of what it may look like earlier this year when a photo of the pope wearing a swagged-out puffy jacket went popular on social media. Though the image was ultimately disproven to be fake, it served as a warning that even modestly sophisticated techniques may propagate harmful disinformation if used to satellite and political photography.
While Meta has released the LLaMA 2 big language model as open source, it has yet to roll out generative AI capabilities for its products like Instagram that are aimed at the general public. However, we have some clues as to the types of things it works on. App researcher Paluzzi has spotted signs of an “AI brush” feature for Instagram that could “add or replace specific parts” of images. In a June all-hands meeting, CEO Mark Zuckerberg said Meta was developing features like using text prompts to modify photos for Instagram Stories. According to the Financial Times, Meta may provide a ‘personas’ function for its AI chatbots as early as next month.
In addition to Meta, Google has unveiled a new tool that will let people tell if a picture is fake. Launching this summer, “About this image” will show users where an image was initially indexed by the search engine, offering useful context for understanding the image’s provenance.
