UNITE AI

Scientists are now tackling the AI problem with AI itself. Researchers from UC Riverside have created a UNITE model to address the grave problem of deepfakes. “People deserve to know whether what they’re seeing is real,” said Rohit Kundu, a PhD candidate from UCR’s Marlan and Rosemary Bourns College of Engineering, who led the paper ‘Towards a Universal Synthetic Video Detector: From Face or Background Manipulations to Fully AI-Generated Content.’ “And as AI gets better at faking reality, we have to get better at revealing the truth.” The researchers have collaborated with scientists from Google, an Alphabet company, to develop a new AI model that detects video tampering and exposes fake content, which is being used to spread disinformation and incite harm. The study noted: “The rapid spread of misinformation, particularly during critical periods such as elections, highlights the need for generalizable detection models capable of identifying diverse manipulations, including face, background, and fully AI-generated T2V/I2V content with/without human subjects.” The model is capable of detecting both partially manipulated and fully synthetic videos. Rather than focusing just on the face, as most conventional detectors do, this model analyzes entire frames, regardless of whether a human subject is present in the videos. This makes it a powerful tool that can be used by fact-checkers, educators, editors, social media platforms, and others to prevent doctored videos from going viral. 

The Rise of AI and the Resulting Synthetic Overload

  Artificial Intelligence (AI) holds tremendous potential in revolutionizing various aspects of both our lives and work. The capability of this technology in automation, data analysis, and decision-making has already begun to transform industries, projected to add multi-trillion dollars to the global economy by the end of this decade. A study by market-forecasting giant IDC estimates that the rise of AI will boost the global economy by a cumulative $19.9 trillion by 2030. McKinsey’s research, meanwhile, expects the value added by generative AI to be as much as $4.4 trillion across the 63 use cases analyzed by the global management consulting firm. About 75% of the value that AI could deliver would be just across four fields: 

 
• R&D

 

• Software Engineering

 

• Marketing and Sales

 

• Customer Operations

 

 While the impact of the technology is forecasted to be significant across all sectors, tech and banking could see the most impact as a percentage of their revenues from gen AI. Goldman Sachs is of the same view, expecting a 7% increase in global GDP from AI. The bank’s economists, Joseph Briggs and Devesh Kodnani, at the time noted: “Despite significant uncertainty around the potential for generative AI, its ability to generate content that is indistinguishable from human-created output and to break down communication barriers between humans and machines reflects a major advancement with potentially large macroeconomic effects.” However, this same capability of the computer system to perform tasks like learning, problem-solving, and decision-making that typically require human intelligence, which is all set to shake up the world, is also sending the world into chaos. The more sophisticated the tech is getting, the blurrier the line between what’s real and what’s not is becoming. 

Why Old Deepfake Detectors No Longer Work

 Over the past few years, advances in AI have led to an unprecedented surge in synthetic media. Estimates suggest that more than half of longer LinkedIn posts are currently written by AI. Then there’s ‘AI slop’, which refers to low-quality, mass-produced AI-generated content. But most concerning of all are deepfakes, which are images, videos, or audio recordings that have been generated or altered using AI. It’s fabricated content that uses AI to present a false representation as realistic. Today, this kind of content is everywhere, permeating all corners of the Internet. These hyper-realistic digital media are causing confusion and spreading misinformation. It is also posing a threat to people’s privacy and security. Cybercriminals are utilizing AI to up their game, conducting phishing scams and identity thefts with alarming precision. According to Kundu: “It’s scary how accessible these tools have become. Anyone with moderate skills can bypass safety filters and generate realistic videos of public figures saying things they never said.” In one such incident, cybercriminals posed as a company’s chief financial officer (CFO) during a Zoom meeting, resulting in a $25 million loss. This is just the beginning, though, as Deloitte predicts that fraud losses from such incidents will hit $40 billion in the US by 2027, up from $12.3 bln in 2023. A US Treasury report has also found that “existing risk management frameworks” adopted by firms “may not be adequate to cover emerging AI technologies.” That’s not to say that there are no tools to help detect AI content and protect oneself against the technology’s risks. There are actually many tools available on the market. The very same companies that are launching new AI tools to make it easy to create new content are also introducing ways to help spot synthetic data. Back in 2020, tech giant Microsoft announced a Video Authenticator to analyze a still photo or video to provide a confidence score in order to determine if the media is artificially manipulated. The tool works by detecting the deepfake’s blending boundary and subtle fading that the human eye may not be able to detect. At the time, it also introduced technology to identify forged content and confirm the authenticity of the media people are interacting with. It included a tool that enables a creator to add digital hashes and certificates to their content, which lives within it as metadata. A reader, meanwhile, was introduced to check the certificates and match the hashes for content authenticity. The tech giant did warn of the tech’s short-term utility in the AI-fueled age. Since deepfakes are generated by AI that continuously learns, it’s only a matter of time before they surpass traditional detection methods. Around the same time, Facebook, a Meta company, also kicked off a competition to develop a deepfake detector using the data that researchers didn’t previously have access to. A few years ago, Intel came up with a FakeCatcher, a real-time deepfake detector that it claims to have an accuracy of 96%. The tool made use of OpenVino to run AI models for face and landmark detection algorithms, while computer vision blocks were optimized with its Integrated Performance Primitives and OpenCV. As for its hardware, the platform can run more than seventy different detection streams at the same time on its 3rd-gen Xeon Scalable processors. Instead of trying to find what’s wrong, FakeCatcher looks for authentic clues by assessing what makes us human and then having algorithms translate those signals into spatiotemporal maps, and finally, using deep learning to instantly detect whether a video is real or fake. Last year, OpenAI also announced that it was researching tools to help with content authenticity. This includes text watermarking, which it noted is effective against localized tampering but not so much against globalized tampering. It also stated that it could “disproportionately impact” groups like non-native English speakers. This update came after the Wall Street Journal reported that the company has already developed a tool that watermarks and detects ChatGPT-generated text with “high accuracy” for some time, but has yet to come to a decision to release it. Additionally, OpenAI has joined the Steering Committee of C2PA (the Coalition for Content Provenance and Authenticity), a widely used standard for digital content certification. The company adds C2PA metadata to all the images created and edited by all of its services, as part of image detection tools. Now, this year, Google also came up with its own AI-generated text, image, audio, and video detection tool called SynthID Detector. The tool from Google, however, only works for content that’s been generated using the tech behemoth’s own AI services like Gemini, Imagen, Veo, and Lyria. This is because the tool basically identifies the presence of a “watermark” that Google’s products have embedded in their output. A watermark is a unique, machine-readable element that’s embedded in content. Unrecognizable by us humans, it can be detected and extracted by algorithms built for that purpose. 

Inside the Tech Powering UNITE’s Breakthrough

  So, as the AI tech advances rapidly, so do the tools to detect the content generated with its help. But there’s no such universal tool that can be used by all on all kinds of AI content. Moreover, the focus of existing deepfake detection techniques, in particular, remains on facial manipulations like lip-syncing or face-swapping, and advancements in tech are rendering them inadequate. With technological innovation making significant progress in text-to-video (T2V) and image-to-video (I2V) generative models, it is now possible for anyone to easily create highly convincing, fully AI-generated synthetic content and seamless background alterations. This, of course, puts everyone from individuals to institutions and even nations at serious risk. Against this background, the complete dependence of earlier deepfake detectors on the face makes them outdated in today’s more technologically advanced world. “If there’s no face in the frame, many detectors simply don’t work. But disinformation can come in many forms. Altering a scene’s background can distort the truth just as easily.” – Kundu So, conventional detectors do not work on newer manipulations, as the new synthetic content now featuring full scenes and backgrounds poses a challenge to face-centric detection methods. This demands a more versatile approach. As a solution to this problem, researchers from UC Riverside have introduced UNITE. The Universal Network for Identifying Tampered and Synthetic Videos (UNITE) model captures full-frame manipulations. “Deepfakes have evolved,” said Kundu, whose focus at UC Riverside is on leveraging foundation models for advanced vision tasks, including image segmentation and fake media detection. “They’re not just about face swaps anymore. People are now creating entirely fake videos – from faces to backgrounds – using powerful generative models. Our system is built to catch all of that.” The model extends detection capabilities to scenarios where there are no faces or no human subjects, and on top of that, it can identify subtle spatial and temporal discrepancies and even cover complex background modifications that previous systems have missed. So, by examining faces as well as background and motion patterns, thereby covering full video frames, UNITE offers one of the first such tools to identify synthetic videos that do not rely merely on facial content. For this, the model utilizes a transformer-based deep learning model, a type of neural network that employs a multi-head attention mechanism to process sequential data. Here, text is converted to numerical representations called tokens. This architecture is actually the foundation for many modern language models like GPT. By processing information in parallel, transformers can facilitate faster training and improved performance. In the case of UNITE, the transformer-based architecture processes domain-agnostic features that are extracted from videos through the SigLIP-So400M foundation model. The foundational AI framework SigLIP extracts features not bound to a specific object or person. Due to limited datasets that cover changes to both facial/background and text-to-video/image-to-video content, the team used innovative training strategies for their model. This means training using data that’s task-irrelevant along with standard deepfake data. So, in addition to the popular FaceForensics++ dataset, the team also used the SAIL-VOS-3D dataset, which simulates complex environments, offering diverse synthetic scenes helpful for training AI detection models. Notably, this was originally designed for 3D video object segmentation in the game GTA-V. While not AI-generated, the dataset is fully synthetic and, as such, helps simulate AI-generated media. This, the team found, enhances their model’s ability to detect various forms of synthetic manipulation. Google provided access to the required datasets as well as computing resources to train the model. In order to reduce the model’s propensity to over-focus on faces, the team also used an attention-diversity (AD) loss, which encourages varied spatial attention throughout video frames. AD loss has been combined with cross-entropy, also known as the log loss function, and commonly used in machine learning (ML) to measure the performance of a classification model, in order to improve the model’s performance across diverse situations. Just training the model on cross-entropy (CE) loss tends to make it hard for it to handle videos where there’s a real human subject with a manipulated background or content generated by T2V and I2V models. Hence, the team introduced AD loss, which prompts the system to monitor multiple visual regions in each frame, thereby boosting its model’s ability to capture important signs from both the foreground and background. AD loss marks the key innovation in the team’s approach, enabling UNITE to not only excel at detecting AI-generated and background-altered videos but also have a noticeable improvement in identifying the usual face-manipulated content. Upon comparing the performance of UNITE with other models, the team found that it “outperforms state-of-the-art detectors on datasets (in crossdata settings) featuring face/background manipulations and fully synthetic T2V/I2V videos, showcasing its adaptability and generalizable detection capabilities.” In a world that’s becoming increasingly digitized and automated, the new system offers a universal detector that can flag a range of fakes, from simple facial swaps to complex, fully synthetic videos created without any real footage. According to Kundu: “It’s one model that handles all these scenarios. That’s what makes it universal.” Currently under progress, UNITE, according to its creators, is a valuable tool in the developing synthetic video detection landscape. Soon, it can be expected to play a key role in defending against video disinformation. 

Investing in AI-based Detection

 In the AI realm, Palantir Technologies is known for its AI-powered data integration, pattern recognition, and anomaly detection. The company operates through four main software platforms: Gotham, Foundry, Apollo, and AIP. Apollo is a single control layer that coordinates configuration, security updates, and delivery of new features to ensure the continuous operation of critical systems. Gotham allows users to identify patterns hidden deep within datasets, while Foundry serves as the operating system for effective asset and risk management. AIP enables firms to run LLMs and other models with full control. 

Palantir Technologies (PLTR )

 Palantir boasts deep ties with the US government, military, and intelligence agencies. This year, it obtained a $30 mln contract to bring AI analysis to immigration records. With a market cap of $372 billion, PLTR shares are currently trading at $157.72, up a whopping 109.35% YTD, thanks to AI demand, massive retail interest, and expanding government contracts. Its EPS (TTM) is 0.23, and the P/E (TTM) is 687.90. 

 Financially, Palantir reported a 39% YoY increase in revenue to $884 million in Q1 2025. Its US revenue, meanwhile, grew 55% YoY to $628 million, including $255 million in US commercial revenue and $373 million in US government revenue. During this period, the company booked its highest quarter of US commercial total contract value, with the remaining deal value at $2.32 billion. Palantir’s customer count in 1Q25 increased by 39% YoY. Its GAAP earnings per share were $0.08, and adjusted EPS was $0.13. Cash, cash equivalents, and short-term US Treasury securities were $5.4 billion at the end of the quarter. “We are delivering the operating system for the modern enterprise in the era of AI. We are in the middle of a tectonic shift in the adoption of our software, particularly in the U.S.” – CEO Alexander C. Karp 

Latest Palantir Technologies (PLTR) Stock News and Developments

Conclusion

 The advent of artificial intelligence has completely changed the world, with both individuals and organizations increasingly embracing the technology to improve productivity and enhance decision-making. Projected to contribute trillions to the world economy, AI isn’t without its peril, though. Deepfakes and their usage to misinform and defraud people are one of the most critical hazards of AI’s widespread adoption. As it becomes harder to differentiate between what’s real and what’s synthetic, tools like UNITE become all the more important and urgent. With this generalizable AI model as the protection against forged content, we may be able to mitigate the negative impact of AI while augmenting and enjoying its positive effects on our work and lives. Click here to learn all about investing in artificial intelligence. 

References:

 1. Kundu, R.; Xiong, H.; Mohanty, V.; Balachandran, A.; Roy‑Chowdhury, A. K.; et al. Towards a Universal Synthetic Video Detector: From Face or Background Manipulations to Fully AI‑Generated Content. arXiv preprint arXiv:2412.12278 (2024). https://doi.org/10.48550/arXiv.2412.12278