Compass Vision: Fighting AI-Generated Deepfakes and Disinformation with Blackbird.AI’s Constellation Narrative Intelligence Platform

The line between reality and fabrication is increasingly blurred. Generative AI has revolutionized content creation, offering unprecedented art, advertising, and entertainment possibilities. However, this technological marvel also harbors a dark side: the potential to spread deepfakes, misinformation, and disinformation with convincing visuals that manipulate public perception and cause massive financial, operational, and reputational harm to organizations across the globe.

AI-generated content can be weaponized to create and amplify harmful narratives, making disinformation more believable and challenging to detect. The damage caused by such deceptive content can profoundly erode trust in media, institutions, and even our judgment. More than simply simply identifying whether the content is AI-generated is required. To fully understand and mitigate the impact, organizations need to gain awareness of the harmful narratives impacting them, the networks those narratives spread across, the threat actors behind the narratives and the bots and automation that scale them, and the hyper-agenda-driven threat actors, cohorts, and communities that are influencing the conversation. Without this understanding, optimal strategic decision-making–especially in times of crisis–becomes impossible.

LEARN MORE: Compass Vision: Instant and Accurate Deepfake Detection

Detection Alone Is Not Enough

Tools that detect AI-generated media are essential but address only a fraction of the problem. Knowing that an image is artificially created doesn’t reveal how it’s being used to influence opinions or behaviors. Disinformation thrives on isolated pieces of false content, the narratives they drive, and the emotional responses they evoke.

Understanding the spread of these narratives is crucial. That involves analyzing how AI-generated content is shared, who promotes it, and what conversations it fuels. Without understanding the full context, efforts to combat disinformation are like addressing only part of the problem. While detecting AI-generated content is valuable, our customers focus on mitigating the impact by understanding the harmful narratives impacting their organization and the threat actors behind them. The goal is to manage the influence of disinformation effectively and reduce risk, even if we can’t fully eliminate its sources. 

How Blackbird.AI Solves the Problem

Enter Blackbird.AI and its Narrative Intelligence Platform Constellation^TM, a system designed to provide a comprehensive solution to the challenges posed by AI-generated misinformation and disinformation. Blackbird.AI’s Constellation^TM Platform surfaces narratives via the Narrative Feed, where AI-generated content is flagged as a critical risk signal. But it doesn’t stop there. We just launched Compass Vision, a new AI-based product that provides instant and accurate analysis and context for identifying AI-generated and AI-manipulated or altered deepfake images and videos. Customers can access Compass Vision directly through Constellation and a Compass Vision API that easily integrates into their existing threat intelligence and social listening systems. Compass Vision is the latest innovative technology that Blackbird.AI has brought to market to offer its customers the most technologically advanced protection against narrative attacks created by misinformation and disinformation. 

LEARN MORE: 8 Ways for Security Leaders to Protect Their Organizations from Mis/Disinformation Attacks

Use Case: Countering the Haitian Migrants Disinformation Campaign in Ohio

In September 2024, an unsubstantiated narrative claiming that Haitian migrants were consuming household pets in Ohio spread across social media, illustrating the complex challenges of modern narrative attacks. 

Blackbird.AI’s Constellation^TM Narrative Intelligence Platform and Compass Vision provide crucial insights into the narrative’s origins, spread, and impact that would empower users with a timely and effective response.

Identification of the Disinformation

Constellation^TM initially detected several AI-generated images and misleading posts that contributed to the narrative alleging harmful activities by Haitian migrants. The platform’s Narrative Feed flagged these narratives as high-risk due to their potential to incite xenophobia and violence and spread disinformation through bots or anomalous behavior.

LEARN MORE: Social Media Misinformation and Disinformation Attack Readiness and Response Checklist

Deeper Analysis Through Network Visualization

The Network Graph visualized within Constellation^TM demonstrates how misinformation and disinformation spread through an echo chamber of agenda-driven influencers and their followers, including significant bot-like activity. In Figure 1 below, the network graph demonstrates the interactions between authors discussing the immigrants consuming pets in the Ohio narrative. The network graph in Figure 1 illustrates how influencers, shown as large central nodes surrounded by clouds of individual accounts engaging with their content, are driving this narrative.  We’ve also applied our bot-like metric to this graph to highlight how a significant proportion of the engagement with these influential accounts comes from bot-like accounts. In Figure 2, we exclusively show the bot-like accounts (shown in dark red), which demonstrate the extent to which the “virality” of these influencers and their content is propagated by bot networks –  indicating an orchestrated effort to manipulate public perception and draw more attention to the inflammatory narrative.

LEARN MORE: TAG Infosphere Report: How Misinformation and Disinformation Represent a New Threat Vector

Additional Context and Debunking

Blackbird.AI’s Compass Context Check adds a layer of analysis by providing real-time background information on the claims and their integrity. Compass leverages an autonomous AI agent that does live searches on the internet for context, filtering by reputable sources. Then, it generates a research synthesis of the type shown in the figure below.  Despite widespread sharing, Compass revealed that the narrative was baseless and debunked by reputable sources, including NPR and Reuters. The lack of credible evidence contrasted sharply with the narrative’s widespread acceptance, spurred by its sensational and dynamic content.

Impact of the Misinformation and Disinformation

Further analysis highlighted the narrative’s broad reach and significant engagement, particularly on posts identified as high-risk. This widespread engagement underscored the narrative’s potential for causing real-world harm, including community unrest and the spread of xenophobia.

This use case underscores the critical role of advanced AI tools in identifying, analyzing, and mitigating the effects of disinformation campaigns. Blackbird.AI’s comprehensive approach—combining AI-generated image detection, narrative tracking, network spread visualization, and contextual analysis—is an essential capability for stakeholders aiming to maintain public trust and societal stability in the face of sophisticated disinformation threats. Adverse effects and risks can be minimized through detailed analysis and timely intervention, helping to safeguard our democratic processes and public discourse.

WATCH: Protect Your Organization from Deepfakes, Misinformation and Disinformation Attacks

The Evolution of Disinformation Tactics

Politics has long been the foundation for new disinformation tactics, especially during election cycles. Strategies honed in the political and military arenas often migrate to the corporate world—a trend we at Blackbird.AI anticipated back in 2017. While it took years for enterprise-targeted disinformation to become commonplace, the rise of AI-generated content dramatically accelerates this disruption.

In the coming months, we expect a significant surge in narratives using AI-generated content to attack enterprises. Adversaries may fabricate visuals depicting, for example, a pharmaceutical company’s vaccine harming children, thereby igniting fear and outrage. It could be a case of “evidence” of a key executive involved in wrongdoing.  Or a video of a conversation or a protest that never actually occurred.  Such compelling imagery makes the false narrative much more persuasive and shareable, amplifying its spread, the initial emotional response, and the potential for harmful impact.  The reason for this is that for many years, the public could trust the evidence delivered in an image or video; the possibility that it was “manufactured” was not a concern.  Now, that concern is front and center.

The Memefication of Disinformation

With AI-generated content, adversaries can craft persuasive content that resonates emotionally. Turning these images into memes creates easily digestible “bite-sized” and highly shareable propaganda pieces. This “memefication” enables disinformation to spread rapidly across social media platforms, reaching vast audiences before it can be challenged or debunked. The example presented in the use case section shows that many memes are funny or outrageously creative. This significantly improves the chances they will be shared with family and friends, even by those who think the base assertion is ridiculous.  This vector for drawing engagement, especially when the AI-manipulated content is blended with authentic content to ground it, is an evolving Tactic/Technique/Procedure (TTP) that we project will be increasingly used as the tools to generate content energize authors online.

The accessibility of generative AI tools lowers the barrier to creating such content, making it feasible for even small groups or individuals to launch large-scale disinformation campaigns. This democratization of content creation poses a significant threat to those tasked with maintaining information integrity.

LEARN MORE: What is a narrative attack?

Key Features of Blackbird.AI’s Constellation^TM Platform:

• Detection of AI-Generated Content with Compass Vision: Designed to detect, aggregate, and prioritize narratives that utilize AI-generated images.
• Narrative Feed for Surfacing, Summarizing, and Prioritizing Narratives for Attention: At the push of a button, the Narrative Feed provides an intrinsically multilingual and multiplatform AI-generated executive summary of top narratives across the conversational landscape. It is inherently multiplatform and multilingual in combining and tracking narratives. Those findings are presented in a risk order instead of just sorted by volume and engagement, and the analyst can adapt the risk calculus to fit the scenario and circumstance on the fly.  
• Composite Risk Calculus: At the heart of Blackbird’s system is a novel risk engine that allows a multi-factor fusion of risk signals.  In the context of this composite risk engine, detecting AI-generated deep fake content is just another signal, albeit often a powerful one.  This scoring can be combined with signals related to propagation and impact, such as bot activity, anomalous propagation patterns, cohort participation, disinformation content, toxicity levels, negative sentiment, and multimodal brand risk (images, videos, audio) to gauge the potential impact on public perception. These signals highlight the most influential and potentially harmful narratives that require attention.  The ability to combine them intuitively and performantly is a patent-pending innovation.
• Network Visualization: This feature allows users to visualize the spread of harmful narratives. It reveals how disinformation spreads through social networks, including synthetic amplification by bots, trolls, and specific cohorts.

Key Features of Blackbird.AI’s Compass Vision

• Simple User Interface: Easily upload images and videos for analysis via a simple portal experience.
• Confidence Score: Deepfake technology is constantly advancing. Compass Vision provides a confidence score with every result to ensure our models evolve as threats evolve.
• Assessment: Compass Vision provides a clear ‘fake’ or ‘not fake’ assessment
• Easy-to-understand Explanation: Our Constellation Narrative Intelligence Platform leads the industry by delivering easy-to-understand contextual observations about AI-generated deepfake images and videos.
• Public API: Easily integrate Compass Vision with existing threat intelligence or social listening systems your team already uses to move faster to determine the authenticity of AI-generated deepfake images and video

By providing an end-to-end solution, Blackbird.AI detects harmful AI-generated content and uncovers the narratives it supports and the mechanisms it proliferates.

WATCH: Go Beyond Social Listening: Using Narrative Intelligence to Counter Mis/Disinformation Attacks

Blackbird.AI: Your Defense Against Narrative Attacks

In this complex and rapidly evolving landscape, Blackbird.AI empowers a robust defense with a deep situational understanding of the spread and impact of AI-generated deepfakes, misinformation, and disinformation. By focusing on the narratives rather than just isolated pieces of content, Blackbird.AI provides a deeper understanding of how disinformation operates and enables you to best protect against it.

Benefits of Using Blackbird.AI:

• Comprehensive Detection: Identifies AI-generated content and assesses their role within larger narratives.
• Unlayers the hidden tactics of modern influence campaigns, revealing the networks and entities involved in spreading disinformation, including bots and troll farms.
• Actionable Intelligence: Equips organizations with the data needed to counteract false narratives and reduce their impact.
• Brand Protection: Helps safeguard reputation by detecting and addressing disinformation before it causes significant harm.

The proliferation of AI-generated content marks a new era in the fight against disinformation. Social listening and threat intelligence detection tools are necessary but need to be improved, and they fall significantly short of their ability to solve this problem. To effectively combat the spread of harmful narratives, we must understand the complex web of factors contributing to their dissemination and influence.

Blackbird.AI rises to this challenge by offering a comprehensive Platform beyond mere detection. Blackbird.AI empowers organizations to proactively defend against these sophisticated attacks by illuminating the narratives and networks behind AI-generated disinformation.

In a world where alternate realities can be fabricated and weaponized, staying ahead of disinformation requires constant innovation and vigilance. Blackbird.AI provides the tools and insights to navigate these new threat vectors, protecting enterprises and the public from the hidden threats lurking in our digital midst. 

LEARN MORE: Blackbird.AI’s Constellation Narrative Intelligence Platform

Integrating Compass Vision into the Defense Against AI-Generated Disinformation

Compass Vision is an important component of Blackbird.AI’s comprehensive strategy to combat AI-generated deepfakes and disinformation. Compass Vision is an advanced AI-based solution specializing in the instant and accurate analysis of images and videos to determine their authenticity. By seamlessly integrating with the Constellation™ Narrative Intelligence Platform, Compass Vision enhances detection capabilities by providing confidence scores and clear assessments for each piece of content analyzed. Its user-friendly interface allows analysts to upload media effortlessly, receiving immediate insights into whether content is AI-generated or manipulated.

Moreover, Compass Vision’s public API enables organizations to integrate this powerful tool into their existing threat intelligence and social listening systems. This interoperability ensures that teams can swiftly authenticate content and make informed decisions in real-time. By adding Compass Vision to their arsenal, organizations strengthen their ability to proactively identify and mitigate the risks associated with AI-generated disinformation, safeguarding their reputation and maintaining public trust.

LEARN MORE: How to Combat Misinformation and Disinformation: Lessons from a Social Media Trust and Safety Experts

The Way Forward – Key Takeaways for Organization Leaders on Navigating Narrative Attacks, Deepfakes, and Disinformation:

Deepfakes Pose Serious Reputational Risks

AI-generated deepfakes can be weaponized to spread false narratives that damage an organization’s reputation. Leaders must be vigilant about the risk that fabricated images or videos pose to public trust and organizational credibility.

Deepfake Detection is Only the First Step

More than simply detecting deepfakes is required. Leaders must understand the broader narratives these manipulated visuals contribute to and the emotional impact they can have on audiences, which may amplify the spread of disinformation.

AI-Generated Content Fuels More Believable Disinformation

Deepfakes increase the believability of disinformation by creating realistic yet false content. Organizations must be prepared that these fabricated visuals can manipulate public opinion and cause long-term harm if not addressed.

Deepfakes Can Incite Social Unrest and Harm Public Perception

As seen in recent disinformation campaigns, deepfakes can stir social and political unrest by spreading fear and xenophobia. Leaders must consider the broader societal impact of deepfake-driven narratives and act to prevent potential harm.

Proactive Defense Builds Resilience Against Future Threats

Blackbird.AI’s platform provides ongoing protection by tracking narrative trends and potential threats. Leaders should invest in narrative intelligence solutions to stay ahead of evolving AI-driven disinformation tactics and ensure long-term resilience.

Looking ahead, it is clear that the accelerating sophistication of generative AI will continue to reshape the misinformation and disinformation landscape in profound ways. As barriers to synthetic media creation are lowered and bad actors adapt their tactics, we will likely see a growing surge of disinformation campaigns including deepfakes targeting all facets of society, from elections and public health to financial markets and brand reputations. 

The new visual analysis capabilities of Blackbird.AI’s Constellation Narrative Intelligence platform and recently announced Compass Vision products are essential in protecting organizations from narrative attacks created by misinformation and disinformation.

To learn more about how Blackbird.AI can help you in these situations, book a demo.