Thursday, 13 March 2025
27.8 C
Singapore
29.6 C
Thailand
20.7 C
Indonesia
27.2 C
Philippines

Meta updates labels for AI-generated images after photographers’ complaints

Meta updated its "Made with AI" labels to "AI info" after photographers complained about mislabeling non-AI content, aiming for better clarity.

Meta is changing its โ€œMade with AIโ€ labels after photographers raised concerns that the company was incorrectly flagging non-AI-generated content. In a recent update, Meta announced it would rephrase the label to โ€œAI infoโ€ because the current wording โ€œwasnโ€™t always aligned with peopleโ€™s expectations and didnโ€™t always provide enough context.โ€

Changes prompted by feedback

Earlier this year, Meta introduced the โ€œMade with AIโ€ labels following criticism from the Oversight Board regarding its โ€œmanipulated mediaโ€ policy. Meta stated it would use โ€œindustry standardโ€ signals to determine when generative AI had been used to create an image. However, it wasn’t long before photographers noticed that Facebook and Instagram were tagging images with this label, even if they had not been made with AI. According to tests by PetaPixel, even minor edits with Adobeโ€™s generative fill tool in Photoshop would trigger the label, such as when editing a โ€œspeck.โ€

While Meta didnโ€™t specifically mention Photoshop in their update, the company noted that โ€œsome content that included minor modifications using AI, such as retouching tools, included industry standard indicators,โ€ which led to the โ€œMade with AIโ€ badge being applied. These indicators could include specific patterns or structures commonly associated with AI-generated content. โ€œWhile we work with companies across the industry to improve the process so our labelling approach better matches our intent, weโ€™re updating the โ€˜Made with AIโ€™ label to โ€˜AI infoโ€™ across our apps, which people can click for more information,โ€ Meta said.

New label, same information

Despite the new label, the โ€œAI infoโ€ tags will not provide specific details about which AI-enabled tools may have been used for the image in question. However, a Meta spokesperson confirmed that the contextual menu that appears when users tap on the badge will stay the same. This menu currently offers a generic description of generative AI and explains that Meta may add the notice โ€œwhen people share content that has AI signals our systems can read.โ€ Users can access this menu by tapping on the ‘AI info’ badge, which will provide additional information about the AI usage in the image.

Moving forward

Metaโ€™s decision to update its AI labels is part of an ongoing effort to align with user expectations and provide a more precise context for its community. As AI continues to play a larger role in media creation and editing, companies like Meta are working to refine their policies and tools to better serve their users. This ongoing process ensures that the community is always part of the conversation.

In the meantime, photographers and content creators should expect the new โ€œAI infoโ€ labels to start appearing on their images. However, with the promise of more precise and informative labelling in the future, Metaโ€™s commitment to improving the accuracy of its labels is a step towards balancing transparency with the practicalities of AI integration in digital media, offering a brighter future for the community.

Hot this week

Endowus partners with Twilio to boost platform security and client trust

Endowus partners with Twilio to enhance security, using advanced verification tools to protect users from fraud while maintaining a seamless experience.

Salesforce to invest US$1 billion in Singapore over five years

Salesforce is investing US$1 billion in Singapore over five years to drive AI innovation, expand workforce development, and enhance local data residency.

Signal president warns AI agents pose serious privacy and security risks

Signal President Meredith Whittaker warns at SXSW that AI agents pose serious risks to privacy and security by requiring deep system access.

Armis acquires OTORIO to enhance on-premises security and strengthen cyber physical systems protection

Armis acquires OTORIO to expand its on-premises cybersecurity solutions, strengthening OT, ICS, and CPS protection for critical industries.

EduSpaze welcomes seven edtech startups in its 10th cohort to transform learning in Southeast Asia

EduSpaze welcomes seven edtech startups to its 10th cohort, focusing on AI-driven learning, job readiness, mental health, and workforce upskilling.

Singapore Airlines and Scoot to ban in-flight power bank charging from April 1

Singapore Airlines and Scoot will ban in-flight power bank use from April 1 due to safety concerns over battery fires. Check their new policies here.

Sandmarc launches 10x optical zoom lens for iPhones, leaving Android users amused

Sandmarc launches a 10x optical zoom lens for iPhones, enhancing long-range photography while amusing Android users already using this feature.

Lego unveils 1,972-piece Mario Kart set with posable arms and head

Lego unveils a 1,972-piece Mario Kart set featuring a posable Mario figure and display stand, which will be available on May 15 for US$249.90.

Trump vows to classify violence against Tesla as domestic terrorism

Trump vows to classify attacks on Tesla dealerships as domestic terrorism, sparking debate over protests, government cuts, and Muskโ€™s influence.

Related Articles