Content Integrity: Dealing with the flood of fabricated content
Content Integrity is one of the five trends identified by the Communications Trend Radar 2024. Advances in synthetic media, and particularly generative AI, have not only increased the amount of manipulated content but have also achieved a disturbing level of realism. This makes it increasingly difficult, if not impossible, to verify the source and authenticity of content.
The escalating risk associated with technology-enabled content generation and modification increases the need to ensure the integrity of the data and content interpreted, used, and produced by communications departments and professionals.
How we make sense of the world
Today, many of our experiences are mediated by digital technologies. These technologies have a profound impact on how we preceive the world, how we gather information, and how we make decision. We’re facing many challenges:
- The advent of synthetic media and especially generative AI, has made it cheaper, easier, and faster to create or alter media content. This can be abused to intentionally mislead recipients.
- Due to the abundance of information online, the time and cognitive resources available for verifying information have decreased. As a result, there is more fabricated or altered media content, and people are less likely to recognize it as such.
- In addition, it has become easier to distribute manipulated or inauthentic content, for example, through the use of bot accounts.
Deepfakes are created by AI algorithms that overlay or replace faces and voices with remarkable realism, making it appear that real people say or do things they never did. Because deepfakes can pose a threat to democracy, security, and personal reputation, identifying them is critical. Some of them can also be detected by
- Facial landmarks and biological signals: Inconsistent head poses, unrealistic or absent eye blinking, variations in facial texture, reflections and details in teeth and eyes.
- Synchronization: Inconsistencies between frames or modalities, for example, between mouth movements and spoken words (phonemeviseme- mismatches).
- Environment: Abnormalities in the background, such as lighting and shadows, or unexpected changes in contrast.
However, as deepfakes continue to improve, it is expected that humans will no longer be able to identify them. This is where media forensic comes in. Media forensics encompasses a range of disciplines aimed at analyzing and verifying multimedia content such as images, audio, and video. It uses techniques from digital signal processing, computer vision, and audio analysis to authenticate content and detect alterations, deep fakes, or manipulations.
Content Integrity refers to the rising need to ensure genuineness of data and material used in communication processes, in the face of growing risks associated with technology-enabled content generation and modification.
Relevance of content integrity for corporate communications
- Communication strategies and activities often rely on the interpretation of information from various sources to assess public debates, stakeholder expectations, media use, outreach of alternative channels, etc. The use of manipulated data and content can lead to misleading advice for top management on communication challenges or ineffective stakeholder communication.
- Communicators use externally and internally sourced material (e.g., survey data, images, videos, animations) to create their own content. Incorporating inaccurate data and content into corporate communications can undermine its effectiveness and create legal and ethical issues.
- Third parties may misuse corporate content by altering text or video for various reasons. These alterations can cast doubt on the position and activities of a company or its executives in order to attack its reputation. They can also misuse a company’s name and authority to legitimize false news or lure audiences to malicious sites.
Dr Markus Talanow, Dussmann Group
» The emergence of AI will lead to a veritable inflation of content coupled with an unprecedented level of personalization. This will also open the door to misuse. Given this, two things will become increasingly crucial for corporate communication: the reliable authentication of corporate messages, and the good old-fashioned trust stakeholders place in the company and its communicators.«
Recommendations for communication leaders and professionals
- Identify changing expectations and new responses of stakeholders to corporate content due to the general rise of fabricated content, especially in social media.
- Implement verification technologies and routines to check the integrity of external and internal content used by communicators for stakeholder communications and internal advising.
- Use authentication technologies when producing your own content to ensure that it can be verified and trusted by audiences. For example, consider adopting C2PA standards to embed content credentials.
- Strengthen your network of trusted sources, train team members, and promote integrity checking as a social practice by publicly sharing how multimedia content is verified and safeguarded by your communication professionals.
- Develop crisis response plans for different threat scenarios, which may include quick response strategies to correct corporate communications based on false information and preemptive warnings about forged corporate content.
About the study
For the fourth year running, a research team from the University of Potsdam (Prof. Dr. Stefan Stieglitz and Sünje Clausen,MSc) and Leipzig University (Prof. Dr. Ansgar Zerfaß and Dr Michelle Wloka) conducted an applied research project to identify trends likely to influence communication practices in the near future. They analyzed hundreds of scientific articles, conference contributions, reports and online publications. These were systematically collected, weighted according to their importance for corporate communication and condensed into five trends. The Communications Trend Radar is funded by the Academic Society for Management & Communication.