ONVIF Will Keep Footage Authentic in the AI Era

In a conversation with a&s Middle East, Leo Levit, Chairman of the Steering Committee of ONVIF, delves into the challenges posed by generative AI to video integrity, the innovative measures ONVIF is implementing to safeguard video evidence, and the future of authentication standards in an evolving security landscape.

Piše: Orhan Hadzagic
E-mail: orhan.hadzagic@asmideast.com

a&s Middle East: How has the rise of generative AI specifically affected video integrity in the security industry?

Levit: Generative AI and the tools that are publicly available enable nearly anyone to alter video or audio footage in a way that makes it almost impossible to see the content alteration. This is leading to an overall general erosion of trust in video and other digital mediums that is spilling over into the security industry.

a&s Middle East: What are some of the most pressing threats posed by generative AI to the authenticity of security video footage?

Levit:  Video is one of the most crucial pieces of evidence used in criminal investigations, court proceedings, and internal investigations within an organization. In many countries, there is a very robust chain of custody process required during investigations and admitting the video as evidence in court. All of that could be rendered useless if we’re unable to trust that the video itself is not an accurate view of the scene. In theory, anything is possible with generative AI. However, here are some examples of specific changes to video that could be commonplace: Changing video overlays might give the wrong information about the date and time of an event. Cutting out specific frames of video from an event removes the event of interest in the video. Altering a face in a scene with that of another person or placing a firearm in an individual’s hand using generative AI.

Securing the video at the earliest point, at the point of video capture from the camera, is key to ensuring the authenticity and trustworthiness of the video footage from camera to court

a&s Middle East: Could you provide an overview of the new ONVIF specification aimed at protecting video integrity? How will this specification enhance current security standards?

Levit:  ONVIF video authentication provides proof that the video has not been edited since it has left the camera sensor. Securing the video at the earliest point, at the point of video capture from the camera, is key to ensuring the authenticity and trustworthiness of the video footage from camera to court.

On a technical level, a signature (or unique digital fingerprint) is created for each video frame, and the frames are then packaged and “signed” using a digital certificate that is unique to that individual camera. This digital key, or signature, allows a supporting viewer (video management client, video player, etc.) to verify that the video data originated directly from the specific camera and has not been tampered with. Specifically, ONVIF defines how this certificate is included in the video stream to be later verified by a compatible client. There are other existing ways of protecting video but most of these are initiated at the point of the VMS and not at the actual source of the video.

Also, very important here is that ONVIF video authentication enables a single, standardized mechanism for video management systems to verify the authenticity of the video it has received.

a&s Middle East: What are the main technical challenges ONVIF faces in ensuring video integrity against manipulation by generative AI tools?

Levit:  As with all standards, adoption may be a challenge. To overcome this and really make video a trusted source of evidence, ONVIF is planning to release the implementation of video authentication as an open-source project. This will also create transparency in the implementation.

ONVIF is planning to release the implementation of video authentication as an open-source project

a&s Middle East: How would these specifications help law enforcement and judicial systems maintain trust in video evidence? Are there any real-world examples that highlight this need?

Levit:  As we discussed before, it’s crucial that public safety officials and corporate security professionals continue to have trust in video evidence. The ability to use a standardized method for automatically verifying the authenticity of video evidence could prove to save time and resources within law enforcement departments responsible for chain of custody requirements.

a&s Middle East: What kind of industry collaboration is required to successfully implement these video integrity standards across different regions and technologies?

Levit:  The success of ONVIF as an organization is a result of many players collaborating for the benefit of our industry, and the industry’s adoption of our standards to provide benefits to end users and system integrators. The ramifications of generative AI on digital video are not only a security industry issue but one that concerns many other industries as well. ONVIF is exploring collaboration opportunities with other organizations that are seeking ways to mitigate these potential risks.

a&s Middle East: In what ways do you anticipate the new specification impacting corporate investigations, particularly in cases involving fraud or internal security breaches?

Levit: The authenticity of the video and its place in maintaining the chain of custody for video evidence is just as important in a corporate setting. This is particularly true when it comes to internal threats, such as those who have access to the organization’s network or other systems that may be used to store video and other evidence.

a&s Middle East: What role can other emerging technologies, like blockchain or AI-based video authentication, play in supporting ONVIF’s goals for video protection?

Levit: We believe that our device-specific method that ONVIF is using for video authentication is an important step toward ensuring the integrity of video.  Other solutions could very well play a role in the future as technologies advance.

a&s Middle East: Given the rapid advancement of generative AI, how does ONVIF plan to keep these standards adaptable to future AI developments?

Levit: From a logistical perspective, ONVIF is hoping to release video authentication as an ONVIF add-on. Unlike an ONVIF profile, which cannot be changed or updated, an add-on can have multiple versions to incorporate new technologies, such as advancements in cybersecurity or other threats from future developments in generative AI.

As an open-source project, the ONVIF video authentication initiative enables a wide community of developers to improve the implementation based on changing trends and conditions.

a&s Middle East: Finally, what do you hope to achieve at Intersec Dubai by discussing video integrity, and what key message would you like to convey to industry stakeholders?

Levit: The topic of video integrity is only beginning to receive wide attention within our industry, but the risks are very real. A standardized solution to this issue would provide assurances to stakeholders that video evidence remains trustworthy and reliable. We aim to highlight the risks and ways to mitigate those risks.

A standardized solution to this issue would provide assurances to stakeholders that video evidence remains trustworthy and reliable