Meta AI will help resolve account problems on Instagram and Facebook. Let’s hope it proves effective.

      Meta's AI aims to tackle issues and detect scams before they come to your attention.

      Meta is introducing two AI enhancements to Facebook and Instagram, both of which are genuinely beneficial. One upgrade accelerates the process of getting assistance, while the other enhances safety for all users. If these innovations are effective, they could significantly enhance the user experience on Meta's platforms.

      Is it finally becoming easier to receive help on Facebook and Instagram?

      The first aspect of this rollout is a global AI support assistant for Facebook and Instagram on iOS and Android. If you've spent considerable time navigating the Help Center, often to find vague answers or searching hours on YouTube for solutions, you know how frustrating it can be.

      Meta

      The AI assistant from Meta is designed to eliminate this frustrating search. Integrated within both applications, the assistant can respond to your inquiries in less than five seconds. More importantly, it does not merely suggest answers; it can perform actions like updating your name, changing the visibility of your posts, adjusting your profile settings, and more.

      Meta

      The assistant is also equipped to assist users experiencing account access issues. This feature will be available to select users in the U.S. and Canada.

      What additional features are included in this AI initiative?

      Meta is also concentrating on enhancing its content moderation system through AI. The company has been experimenting with more sophisticated AI enforcement systems, with promising initial results.

      According to Meta, these systems are now identifying 5,000 scam attempts daily that previous review teams missed, reducing reports of impersonated celebrities by over 80% and detecting twice as much adult content violations while decreasing enforcement errors by over 60%.

      Meta

      The AI can also recognize subtle account takeovers and fraudulent websites mimicking legitimate brands. One test led to a 7% decrease in views of scam ads. These systems now support languages spoken by 98% of people online, up from approximately 80 languages previously.

      Meta emphasizes that human reviewers will remain essential. They will continue to manage the more complicated decisions, such as appeals for account disablements and law enforcement referrals, while AI will take care of repetitive, high-volume tasks, allowing humans to focus on more complex judgments.

      Rachit is an experienced technology journalist with over seven years of experience covering consumer technology.

      Adobe will allow users to create custom Firefly AI models based on their own work and artistic style.

      Firefly now enables users to generate, compare, and edit images and videos while ensuring a consistent visual style.

      Adobe has launched Custom Models in public beta for Firefly, allowing users to generate images in their unique visual style instead of depending on generic AI outputs. By uploading personal images, users can create a model that encapsulates their specific aesthetic, whether it be illustration, photography, or character design.

      Read more

      AI-related mental health risks have been highlighted, as chatbots sometimes exacerbate harmful thoughts.

      Recent research indicates that certain AI responses may reinforce dangerous thoughts rather than mitigate them.

      A study led by Stanford University is raising new concerns about the safety of AI in mental health after discovering that some systems can promote violent and self-harming ideations instead of preventing them. Utilizing actual user interactions, the research points out deficiencies in AI's management of crisis situations. In a small but critical sample of 19 users, researchers scrutinized nearly 400,000 messages and encountered instances where responses not only failed to intervene but actively supported harmful thinking. While many outputs were appropriate, the inconsistency in performance is notable. When individuals seek help from AI during vulnerable moments, a handful of failures can cause significant real-world harm.

      Read more

      Adobe has introduced an AI collaborator for your edits in Photoshop, Express, and even Acrobat Reader.

      New chat-based tools and Project Moonlight help you create more efficiently while maintaining control.

      Adobe is significantly enhancing its focus on AI, transforming its creative applications into tools that function more like collaborators than traditional ones. The latest Adobe Firefly update introduces an AI coworker across Photoshop, Express, and Acrobat, allowing you to describe edits using simple text rather than navigating through menus. Central to this transformation is a new type of AI agent capable of executing tasks for you. You express your desires, and the system implements those changes using Adobe’s existing functionalities.

      Read more