Artificial intelligence is making online fraud more convincing than ever. Across the world, cybercriminals are using AI to create fake identities, cloned voices, realistic images and manipulated videos that are increasingly difficult to distinguish from genuine content.
Today’s image-generation models can produce convincing identity cards, fake photos of family members, fabricated evidence and political propaganda within minutes.
As these tools become more accessible, concerns over misinformation, digital trust and AI-enabled fraud also increase.
To address this, technology companies are moving beyond traditional AI detection tools. Instead of relying solely on software that attempts to identify AI-generated content, operating system developers and chipmakers are building verification technologies directly into their platforms.
This approach differs from invisible watermarking systems such as Stable Signature or Google’s SynthID. Rather than depending only on hidden watermarks added to AI-generated files, these newer technologies allow operating systems and hardware to verify the origin and authenticity of supported content using cryptographic signatures and trusted provenance data.
In practical terms, your phone, computer or processor can help confirm where supported images, videos or documents came from and whether they have been edited or created using AI.
Android: Built-in Content Verification
Google is gradually expanding Android’s ability to verify digital content at the operating system level instead of leaving the task entirely to third-party applications.
As AI-generated images, videos and documents become more common, Android is being designed to support industry standards that record where digital content originated and whether it has been altered. This gives compatible apps a trusted way to verify content without creating separate verification systems.
Rather than analysing pixels to guess whether an image was generated by AI, Android focuses on reading Content Credentials and cryptographic signatures attached to supported files.
When an image or video includes verified provenance information, compatible Android apps can display details such as where the content originated, whether it has been edited and whether generative AI tools were involved in its creation.
Because this verification is handled by the operating system, it offers stronger protection than software-only approaches.
Apple: iPhone and Mac Verification
Apple is taking a similar direction through its Apple Intelligence platform and security architecture across iOS and macOS.
Instead of relying on AI models to detect manipulated media, Apple is focusing on preserving and verifying the history of supported files.
This is made possible through Content Credentials, an industry standard that stores cryptographically signed information about a file’s origin and editing history. When supported images or videos are opened on an iPhone, iPad or Mac, compatible applications can verify this information and indicate whether AI tools were used during creation or editing.
Because the verification process is backed by Apple’s secure hardware and software, it provides a higher level of trust than ordinary file metadata, which can be altered.
Windows 11: Verifying Authenticity
Microsoft is also strengthening content verification in Windows 11 through its support for Project Provenance and the Content Credentials standard.
Rather than introducing another AI detection tool, Microsoft is enabling Windows to verify digital content using trusted metadata and cryptographic signatures. This allows supported applications to confirm where media originated and whether it has been modified.
When compatible images or videos contain verified provenance information, Windows can validate that information through its built-in security features. Applications can then display details about the content’s origin, editing history and AI involvement without relying solely on AI detection systems.
Snapdragon and Intel: Security at the Chip Level
Content verification is also moving into computer hardware.
Qualcomm’s Snapdragon platforms and Intel’s latest processors include dedicated security components that establish a hardware “root of trust”. These secure environments store cryptographic keys and perform verification without exposing sensitive information to the rest of the system.
When an operating system or application needs to verify an image, video or document, it can use these hardware security features to confirm that digital signatures and Content Credentials are genuine. Since the process takes place inside protected hardware rather than standard software, it is significantly more difficult for malware or attackers to interfere with the verification process.
Google’s SynthID
Google’s SynthID remains one of the leading software-based approaches to identifying AI-generated content.
Unlike system-level verification, SynthID embeds an invisible watermark into AI-generated content, making it possible to determine whether supported images were created using Google’s AI models.
The technology currently supports content produced by Google’s AI tools, including Imagen, Veo and Nano Banana, while coverage continues to expand to video, audio and some text outputs.
Several technology companies, including OpenAI, ElevenLabs, Kakao and NVIDIA, have also adopted or integrated aspects of SynthID’s watermarking technology into parts of their services, helping broaden its adoption across the AI ecosystem.
The Bigger Picture
What AI can do keeps improving, so do the methods used by fraudsters to deceive people online.
The industry’s response is moving from software-only detection to a bigger trust framework that combines operating systems, processors, secure hardware and cryptographic verification.
These technologies will not eliminate AI-enabled fraud overnight. However, combining system-level verification with watermarking and trusted content credentials gives users a stronger way to assess whether digital content can be trusted before acting on it.
