The feature works by allowing users to upload a selfie-style facial scan, which YouTube then uses to search for potentially manipulated or synthetic videos across the platform. If a match is detected, users receive an alert and can request removal of the content.

The company said the expansion marks a major step in giving ordinary users the same protections previously reserved for public figures, politicians, journalists, and creators.

“With this expansion, we’re making clear that whether creators have been uploading to YouTube for a decade or are just starting, they’ll have access to the same level of protection,” YouTube spokesperson Jack Malon was quoted as saying by The Verge.

The likeness detection system was first introduced in October 2025 and operates similarly to YouTube’s Content ID technology, but instead of scanning for copyrighted material, it searches for a user’s facial likeness.

Users can access the feature through YouTube Studio by navigating to the Content Detection section and selecting the “Likeness” option.

YouTube cautioned that the system may also surface legitimate videos featuring a user’s real face, including clips from their own uploads, which may not qualify for removal under platform policies.

The post YouTube Introduces AI Deepfake Detection Tool to All Adult Users appeared first on Tech Digest News.

According to the company’s status page, the chatbot recorded a “major outage” lasting seven hours and 13 minutes, from the early hours of Monday until 10:33 a.m. local time, when services were restored.

The company did not disclose the cause of the disruption, in line with its standard protocol. However, outages of this nature are typically linked to server malfunctions or software bugs introduced during updates.

While DeepSeek’s API services—primarily used by developers—previously experienced extended outages in January 2025, this marks the first time its main chatbot interface for everyday users has been down for more than two hours.

The incident comes at a time when the global artificial intelligence industry is closely watching the company for the release of its next-generation model. DeepSeek has yet to announce a timeline for the anticipated update.

The post DeepSeek Suffers Longest Outage Since Rise of AI Models appeared first on Tech Digest News.

The decision marks a departure from standard industry practice, where AI developers typically share pre-release versions of major models with leading chipmakers to ensure optimal performance on widely used hardware.

Instead, DeepSeek granted early access to domestic suppliers, including Huawei Technologies, the sources said. The move gives Chinese chipmakers several weeks to optimise the software for their processors ahead of the model’s expected release.

Nvidia and AMD declined to comment, while DeepSeek and Huawei did not respond to requests for comment.

The reasons behind the decision were not immediately clear. However, industry analysts suggest the move could align with broader Chinese government efforts to reduce reliance on U.S. semiconductor technology.

The development comes amid heightened scrutiny over U.S. export controls. A senior Trump administration official recently told Reuters that DeepSeek’s latest AI model was trained using Nvidia’s most advanced Blackwell chips in mainland China, a potential violation of U.S. export restrictions. U.S. policy currently bars shipments of certain advanced AI chips to China.

DeepSeek’s models have been downloaded more than 75 million times on the open-source platform Hugging Face since the company gained global attention in early 2025. Chinese open-source AI models have seen rapid adoption, surpassing downloads of models from other countries on the platform over the past year.

U.S. authorities previously allowed Nvidia’s H20 and AMD’s MI308 chips, designed for AI inference, to resume shipments to China, though licenses for more advanced processors remain restricted. It is unclear whether DeepSeek has obtained approval to purchase those chips.

Analysts say the immediate commercial impact on Nvidia and AMD may be limited, as DeepSeek’s models primarily serve as benchmarks rather than enterprise deployment standards. However, the decision underscores growing technological and geopolitical tensions as Chinese AI firms prepare to release new models in the coming months.

The post DeepSeek Withholds V4 Model From Nvidia, AMD Ahead of Launch appeared first on Tech Digest News.

The official said the U.S. believes DeepSeek may have clustered the advanced chips in a data center located in Inner Mongolia and could attempt to remove technical indicators revealing their origin.

The Blackwell chips are produced by Nvidia and are currently barred from export to China under U.S. Commerce Department regulations.

The official declined to explain how the U.S. obtained the intelligence or how DeepSeek allegedly acquired the chips but emphasized that U.S. policy prohibits shipping Blackwell processors to China.

Nvidia declined to comment, while the U.S. Commerce Department and DeepSeek did not respond to requests for comment.

The Chinese embassy in Washington criticized what it described as the politicization of trade and technology issues, saying Beijing opposes the “overstretching” of national security concerns and expansive export controls.

The allegation could deepen divisions in Washington over how strictly to limit China’s access to advanced AI semiconductors. Some policymakers argue that restricting exports protects U.S. technological dominance and national security, while others warn that aggressive curbs could push Chinese firms to accelerate domestic chip development.

The controversy underscores the intensifying global competition over advanced AI hardware — widely regarded as a strategic asset in both commercial and military domains.

The post U.S. Probes DeepSeek Over Alleged Use of Nvidia Blackwell AI Chips appeared first on Tech Digest News.

In the memo submitted to the U.S. House Select Committee on Strategic Competition between the United States and the Chinese Communist Party, OpenAI alleged that individuals linked to DeepSeek sought to circumvent access restrictions to obtain outputs from U.S. AI systems.

“We have observed accounts associated with DeepSeek employees developing methods to circumvent OpenAI’s access restrictions and access models through obfuscated third-party routers and other ways that mask their source,” the company said.

OpenAI added that DeepSeek employees developed code to access U.S. AI models and collect outputs for use in distillation, a technique in which a more powerful model is used to evaluate and refine a newer model’s responses, effectively transferring knowledge.

DeepSeek and its parent company, High-Flyer, did not immediately respond to requests for comment.

The Hangzhou-based startup drew global attention last year after launching models that rivaled leading U.S. systems, raising concerns in Washington that China could narrow the AI gap despite export controls and restrictions.

OpenAI said it proactively removes users suspected of attempting to distill its models and accused some Chinese developers of cutting corners in safety practices.

Silicon Valley executives have previously praised DeepSeek’s models, including DeepSeek-V3 and DeepSeek-R1, which are available internationally.

The post OpenAI Alleges Chinese Startup DeepSeek Attempted to Clone U.S. AI Models appeared first on Tech Digest News.

The company said V3.2 delivers near state-of-the-art results across coding, tool use and other benchmark tasks, while the Speciale variant recorded gold-medal scores at the 2025 International Math Olympiad and Informatics Olympiad, underscoring its technical depth. Both models remain available under an open-source licence, a key strategy behind DeepSeek’s rapid global adoption.

DeepSeek said V3.2 is powered by three major innovations: DeepSeek Sparse Attention (DSA), a mechanism designed to cut computational costs while preserving performance; a scalable reinforcement learning framework; and a large-scale agentic task synthesis pipeline. The DSA system, which splits attention into two components, is optimised for long-context processing and is the only architectural adjustment added during continued pretraining.

The models are built on the DeepSeek-V3 Mixture-of-Experts transformer, featuring 671 billion total parameters and 37 billion active parameters per token.

DeepSeek also introduced updates to its chat template, including a redesigned tool-calling format and a new “thinking with tools” feature aimed at improving reasoning accuracy.

The startup rose to prominence earlier in the year after the release of its DeepSeek-V3 and DeepSeek-R1 models, which drew global attention for achieving performance close to OpenAI’s latest systems while remaining fully open-source — a rarity among leading AI developers.

The post DeepSeek Unveils V3.2 Models Rivaling GPT-5 and Gemini 3 Pro appeared first on Tech Digest News.

The company announced the model Monday on Hugging Face, alongside a linked research paper on GitHub. V3.2-exp introduces DeepSeek Sparse Attention, which uses a “lightning indexer” to prioritize key excerpts from a large context window and a “fine-grained token selection system” to further filter relevant tokens. Together, these modules enable long-context processing with lower server demands.

Early testing suggests the model could cut the cost of API calls in half for extended-context use cases. Since V3.2-exp is open-weight and freely available, independent researchers are expected to validate those claims soon.

DeepSeek’s work comes amid industry-wide efforts to lower inference costs—the ongoing expense of running pre-trained AI models, which remains a major barrier to scaling AI services. The company says its approach shows the transformer architecture still has untapped efficiency gains.

The Beijing-based startup drew global attention earlier this year with its R1 model, trained primarily with reinforcement learning at a fraction of U.S. competitors’ budgets. While R1 fell short of predictions that it would reshape training economics, DeepSeek’s latest release could influence how providers tackle the cost problem.

Although less headline-grabbing than R1, the sparse attention technique may offer U.S. and global firms new methods to reduce infrastructure strain and improve affordability in large-scale AI deployment.

The post China’s DeepSeek Unveils AI Model With Lower Inference Costs appeared first on Tech Digest News.

The move underscores Beijing’s determination to build a self-reliant semiconductor ecosystem in response to U.S. export restrictions that have limited access to advanced Western technologies.

DeepSeek, which jolted the global AI industry earlier this year by releasing cost-effective models rivaling Western systems like OpenAI’s ChatGPT, said V3.1 features UE8M0 FP8 precision optimization, enabling more efficient use of memory while increasing speed. FP8—or 8-bit floating point—is widely seen as the next step in reducing computing costs for large-scale AI applications.

While the company did not reveal which Chinese-made chips would be supported, it hinted that the model had been tuned for “soon-to-be-released next-generation domestic chips.”

V3.1 also introduces a hybrid inference structure that lets users toggle between reasoning and non-reasoning modes via a new “deep thinking” button on DeepSeek’s app and web platform. Developers using its API will see a revised cost structure beginning September 6.

The update comes after two other significant releases this year: an R1 model update in May and an earlier V3 enhancement in March. Analysts say the focus on local chip support positions DeepSeek as a key player in China’s long-term plan to reduce reliance on foreign AI infrastructure.

The post Amid U.S.-China Chip War, AI Startup DeepSeek Launches Model Optimized for China-Made Chips appeared first on Tech Digest News.

The rise of AI chatbots like DeepSeek and ChatGPT has triggered a wave of unease worldwide. Governments are grappling with concerns about data security and the privacy implications of using these generative AI services.

The (Seoul) South Korea’s state-run Korea Hydro & Nuclear Power has recently blocked access to DeepSeek and other generative AI services. As per the reports, the ministry of other nations has also banned the use of AI services on computers connected to external networks.

The Australian government has banned DeepSeek from all its government systems and devices, citing its use as a national security risk. Tony Burke, the Australian Home Affairs Minister, highlighted that the government has made this decision to protect Australia’s national security.

Recently, Taiwan has also restricted its government agencies from using DeepSeek. Taiwan’s Ministry of Digital Affairs has announced that “all government agencies and critical infrastructure should refrain from using DeepSeek, citing threats to national information security.”

France and Italy including other European nations have restricted DeepSeek to protect their consumers’ data as it involves the risk of data collection as per its privacy policies. Meanwhile, the USA has banned the use of DeepSeek in both its government and private sectors.

The governmental bodies in both the UK and Ireland have raised alarms over the potential security of using the Chinese AI model. However, despite its low-cost service the security risks of the Chinese AI firm have raised concern globally, hence it will be a challenge for DeepSeek.

The post DeepSeek Banned Globally by Different Countries appeared first on Tech Digest News.

The release of OpenAI’s GPT models, particularly ChatGPT and GPT-3, sent shockwaves across the globe. These systems, with their ability to generate human-like text and perform tasks previously considered the exclusive domain of human intelligence, astounded observers.

Prior to ChatGPT, AI was largely confined to narrow applications such as basic chatbots or manufacturing automation. OpenAI’s breakthrough, however, revolutionized the AI landscape. ChatGPT, with its ability to engage in conversations, compose essays, write poetry, and even debug code, demonstrated an unprecedented level of fluency and adaptability. Users, initially expecting responses akin to earlier, more rudimentary chatbots, were often left awestruck by the system’s insightful and human-like output.

Social media platforms quickly became inundated with examples of ChatGPT’s prowess within days of its release. Users shared screenshots of its detailed answers, creative writing, and even humorous outputs, often accompanied by comments like “Is this real?” or “I can’t believe AI knows this!”. For many, this was their first encounter with AI that truly seemed to “think” in a natural and human-like manner.

Overnight, individuals with no prior technical background found themselves wielding powerful AI tools. The experience of interacting with ChatGPT felt akin to having a personal genius assistant, empowering users with newfound capabilities.

The subsequent release of GPT-4, with its multimodal capabilities (processing both text and images) and enhanced features inherited from its predecessors, further solidified the perception of having reached the pinnacle of AI advancement.

It is, therefore, unsurprising that Image Merchants Promotion Limited (IMPR), publishers of PRNigeria and Economic Confidential, during their three-day staff retreat at the PRNigeria Centre Kano, announced their commitment to leveraging Artificial Intelligence (AI) and emerging technologies to enhance productivity in 2025.

In 2024, the organization integrated AI to improve the efficiency and accuracy of its operations, particularly in in-depth special reports and fact-checking.

Mr. Yushau Shuaib, Chief of IMPR, emphasized the need for companies in the strategic communication and public relations industry to fully embrace AI solutions. He stressed the importance of combining human intelligence and creativity with AI to enhance staff productivity and maintain a competitive edge in the face of stiff competition among tech giants.

It was during this retreat that we delved into the apparent rivalry between the West and the East in the realm of Artificial Intelligence. This discussion was particularly fueled by the recent launch of “DeepSeek-V3” by Chinese tech firms, a groundbreaking AI model that has set a new global benchmark for AI applications with its affordability, exceptional performance, and open-source development strategies.

DeepSeek-V3 has taken the AI community by storm. Its performance surpasses many top-tier AI models, including closed-source AI, challenging the notion that open-source AI could only play a secondary role.

DeepSeek-V3 boasts an impressive speed and efficiency, processing information at a blistering 60 tokens per second – a threefold increase over its predecessor.
The model employs a “Mixture-of-Experts (MoE)” architecture. This architecture comprises multiple neural networks, each optimized for specific tasks. When DeepSeek-V3 receives a prompt, a “router” intelligently directs the request to the neural network best suited to handle it. Each of these individual neural networks within the MoE structure possesses 34 billion parameters.

To illustrate, imagine a classroom where each student specializes in a different subject, such as mathematics, storytelling, or art. When a problem arises, the teacher (router) determines which student (neural network) is best equipped to address it. This architecture ensures that users receive the most relevant and effective responses.

Furthermore, DeepSeek-V3 was trained on a massive dataset of 14.8 trillion tokens. In data science, tokens represent units of raw data, with 1 million tokens roughly equivalent to 750,000 words. This translates to the model being trained on an astonishing 11 trillion words.

Prior to DeepSeek-V3, GPT-4 held the record for the largest training dataset with approximately 1 trillion tokens. Considering GPT-4’s impressive performance with that amount of data, one can only imagine the capabilities of DeepSeek-V3.

Another remarkable aspect of DeepSeek-V3 is its remarkably low production cost. The company revealed that they spent only $5.5 million to train the model, a figure significantly lower than the costs associated with developing other leading models, such as GPT-4, which reportedly cost over $100 million.
Moreover, DeepSeek-V3 outperforms other AI models across various benchmarks and evaluation metrics commonly used to assess AI performance.

Reports indicate that DeepSeek-V3, with its 6.71 trillion parameters, surpasses Meta’s Llama 3.1 and outperforms mainstream closed-source models like GPT-4 in numerous benchmark tests. This achievement not only signifies a breakthrough in Chinese AI technology but also represents a significant innovation in the global AI landscape.

The realization that these powerful features are not only accessible but also entirely free to use is truly extraordinary. It underscores the immense potential that can be unlocked without any financial barriers.

The United States and China are undoubtedly the leading forces in AI development, with the US currently holding a slight technological edge. In October 2022, the US implemented export controls to restrict China’s access to advanced NVIDIA AI chips – specialized hardware crucial for accelerating computationally intensive tasks such as deep learning, natural language processing, and computer vision. This move aimed to hinder China’s AI progress and reduce competition in AI technology advancement.

Despite these restrictions, China has demonstrated its ability to overcome obstacles and achieve groundbreaking innovations. The development of DeepSeek-V3 not only defies expectations but also serves as a potential model for the US in advancing and refining its own domestically-driven AI systems. It exemplifies the adage, “What doesn’t kill you makes you stronger.”

The development of DeepSeek-V3 marks a pivotal moment in the evolution of AI technology. It challenges the conventional norms established by previous models like GPT-4 and demonstrates that cutting-edge AI does not have to be prohibitively expensive. DeepSeek-V3 proves that innovation can be achieved while maintaining affordability, setting a new standard for future AI development that emphasizes accessibility and affordability.

The bar has been raised. Now, it is incumbent upon innovators, researchers, and developers to collaborate, compete, and strive to push the boundaries of AI even further.

The future of AI will be shaped by collaboration, healthy competition, and a shared commitment to creating even better and more powerful systems. Together, we can shape the next generation of AI that will define the future.

Shuaib S. Agaka, Tech Journalist, writes from PRNigeria Centre Kano.

The post US-China AI Rivalry Intensifies as DeepSeek Enters the Fray, By Shuaib S. Agaka appeared first on Tech Digest News.