Digital business models

DeepSeek unveils new approach to improve AI reasoning

Chinese AI firm DeepSeek has unveiled a new method to improve LLM reasoning skills, claiming it offers more accurate and faster responses than current technologies. The approach, developed with researchers from Tsinghua University, combines generative reward modeling (GRM) with a self-principled critique tuning technique.

The method aims to refine how AI LLMs respond to general queries by better aligning their outputs with human preferences. According to a paper published on the arXiv scientific repository, the resulting DeepSeek-GRM models showed stronger performance than existing methods and proved competitive against widely accepted public reward models.

DeepSeek has announced intentions to release these models as open source, though no release date has been set. The move follows increased global interest in the company, which had earlier gained attention for its V3 foundation model and R1 reasoning model.

For more information on these topics, visit diplomacy.edu.

Digital Morocco 2030 strategy focuses on tech-driven transformation

Morocco has set ambitious goals to boost its economy through investment in emerging technologies, aiming for a 10% increase in GDP by 2030. As part of its Digital Morocco 2030 strategy, the government is committing over 11 billion dirhams ($1.1 billion) by 2026 to drive digital transformation, create more than 240,000 jobs, and train 100,000 young people annually in digital skills.

The roadmap prioritises digitising government services through a Unified Administrative Services Portal, with the long-term goal of placing Morocco among the world’s top 50 tech nations. Blockchain plays a central role in this vision, being adopted to improve transparency and efficiency in public services, and already undergoing trials in private sectors like healthcare and finance.

Despite an ongoing official ban, digital asset ownership has surged, more than six million Moroccans now hold such assets, representing over 15% of the population. In parallel, the country is rapidly expanding its use of AI. Notably, Morocco has introduced AI into its judiciary, launched an AI-powered university learning system, and trained over 1,000 small- and medium-sized businesses in AI adoption through partnerships with LinkedIn and the European Bank for Reconstruction and Development.

For more information on these topics, visit diplomacy.edu.

Metro Bank teams up with Ask Silver to fight fraud

Metro Bank has introduced an AI-powered scam detection tool, becoming the first UK bank to offer customers instant scam checks through a simple WhatsApp service.

Developed in partnership with Ask Silver, the Scam Checker allows users to upload images or screenshots of suspicious emails, websites, or documents for rapid analysis and safety advice.

The tool is free for personal and business customers, who receive alerts if the communication is flagged as fraudulent. Ask Silver’s technology not only identifies potential scams but also automatically reports them to relevant authorities.

The company was founded after one of the co-founders’ family members lost £150,000 to a scam, fuelling its mission to prevent similar crimes.

The launch comes amid a surge in impersonation scams across the United Kingdom, with over £1 billion lost to fraud in 2023. Metro Bank’s head of fraud, Baz Thompson, said the tool helps counter tactics that rely on urgency and pressure.

Customers are also reminded that the bank will never request sensitive information or press them to act quickly via emails or texts.

For more information on these topics, visit diplomacy.edu.

Trump administration pushes for pro-AI shift in US federal agencies

The White House announced on Monday a shift in how US federal agencies will approach AI, prioritising innovation over the stricter regulatory framework previously established under President Biden. 

A new memorandum from the Office of Management and Budget instructs agencies to appoint chief AI officers and craft policies to expand the use of AI technologies across government operations.

This pivot includes repealing two Biden-era directives emphasising transparency and safeguards against AI misuse. 

The earlier rules required federal agencies to implement protective measures for civil rights and limit unchecked acquisition of AI tools. 

These protections have now been replaced with a call for a more ‘forward-leaning and pro-innovation’ stance, removing what the current administration views as excessive bureaucratic constraints.

Federal agencies are now expected to develop AI strategies within six months. These plans must identify barriers to responsible AI implementation and improve how the technology is used enterprise-wide. 

The administration also encouraged the development of specific policies for generative AI, emphasising maximising the use of American-made solutions and enhancing interoperability between systems.

The policy change is part of President Trump’s broader rollback of previous AI governance, including his earlier revocation of a 2023 executive order signed by Biden that required developers to disclose sensitive training data. 

The new framework aims to streamline AI procurement processes and eliminate what the administration labels unnecessary reporting burdens while still maintaining basic privacy protections.

Federal agencies have already begun integrating AI into their operations. The Federal Aviation Administration, for example, has applied machine learning to analyse safety reports and identify emerging aviation risks. 

Under the new guidelines, such initiatives are expected to accelerate, signalling a broader federal embrace of AI across sectors.

For more information on these topics, visit diplomacy.edu.

SandboxAQ attracts major investors for AI expansion

SandboxAQ has secured a further $150 million in funding, bringing its Series E total to over $450 million.

The quantum-AI firm, which originated at Alphabet in 2016 and spun out in 2022, continues to draw backing from tech heavyweights and financial leaders, including Google, Nvidia, Ray Dalio, and BNP Paribas.

The funding is set to accelerate development in large quantitative models (LQMs), which lie at the heart of SandboxAQ’s enterprise solutions.

The models are already being applied across sectors such as life sciences, finance, and navigation. BNP Paribas described the partnership as a key move at the intersection of AI and quantum technology, while Ray Dalio cited his confidence in the company’s team and approach.

Recent collaborations have also strengthened SandboxAQ’s position in the field. In early 2025, it partnered with Google Cloud to deliver its LQMs via the Google Cloud Marketplace, easing deployment for enterprise users.

Previous deals include a November 2023 alliance with Nvidia to simulate chemical reactions for new material development. In total, SandboxAQ has now raised more than $950 million, achieving a pre-money valuation of $5.3 billion in late 2024.

For more information on these topics, visit diplomacy.edu.

Senator Warner warns TikTok deal deadline extension breaks the law

Senator Mark Warner, the top Democrat on the Senate Intelligence Committee, has criticised President Donald Trump’s recent move to extend the deadline for ByteDance to divest TikTok’s US operations. 

Warner argued that the 75-day extension violates the law passed in 2024, which mandates a complete separation between TikTok’s American entity and its Chinese parent company due to national security concerns.

The deal currently under consideration would allow ByteDance to retain a significant equity stake and maintain an operational role in the new US-based company. 

According to Warner, this arrangement fails to satisfy the legal requirement of eliminating Chinese influence over TikTok’s US operations. 

He emphasised that any legitimate divestiture must include a complete technological and organisational break, preventing ByteDance from accessing user data or source code.

The White House and TikTok have not issued statements in response to Warner’s criticism. In its second term, Trump’s administration has stated it is in contact with four groups regarding a potential TikTok acquisition. 

However, no agreement has been finalised, and China has yet to publicly support a sale of TikTok’s US assets, one of the primary obstacles to completing the deal.

Under the 2024 law, ByteDance was required to divest TikTok’s US business by 19 January or face a ban. 

Trump, who retook office on 20 January, chose not to enforce the ban immediately and instead signed an executive order extending the deadline. 

The Justice Department further complicated the issue when it told Apple and Google that the law would not be enforced, allowing the app to remain available for download.

As the deadline extension continues to stir controversy, lawmakers like Warner insist that national security and legislative integrity are at stake.

For more information on these topics, visit diplomacy.edu.

Osney Capital invests in the UK’s cybersecurity innovation

Osney Capital has launched the UK’s first specialist cybersecurity seed fund, focused on investing in promising cybersecurity startups at the Pre-Seed and Seed stages.

The fund, which raised more than its initial £50 million target, will write cheques between £250k and £2.5 million and has the capacity for follow-on investments in Series A rounds.

Led by Adam Cragg, Josh Walter, and Paul Wilkes, the Osney Capital team brings decades of experience in cybersecurity and early-stage investing. Instead of relying on generalist investors, the fund will offer tailored support to early-stage companies, addressing the unique challenges in the cybersecurity sector.

The UK cybersecurity industry has grown to £13.2 billion in 2025, driven by complex cyber threats, regulatory pressures, and the rapid adoption of AI. The fund aims to capitalise on this growth, tapping into the strong talent pipeline boosted by UK universities and specialised cybersecurity programs.

Supported by cornerstone investments from the British Business Bank and accredited by the UK’s National Security Strategic Investment Fund, Osney Capital’s mission is to back the next generation of cybersecurity founders and help them scale globally competitive businesses.

For more information on these topics, visit diplomacy.edu.

Sam Altman’s AI cricket post fuels India speculation

A seemingly light-hearted social media post by OpenAI CEO Sam Altman has stirred a wave of curiosity and scepticism in India. Altman shared an AI-generated anime image of himself as a cricket player dressed in an Indian jersey, which quickly went viral among Indian users.

While some saw it as a fun gesture, others questioned the timing and motives, speculating whether it was part of a broader strategy to woo Indian audiences. This isn’t the first time Altman has publicly praised India.

In recent weeks, he lauded the country’s rapid adoption of AI technology, calling it ‘amazing to watch’ and even said it was outpacing the rest of the world. His comments marked a shift from a more dismissive stance during a 2023 visit when he doubted India’s potential to compete with OpenAI’s large-scale models.

However, during his return visit in February 2025, he expressed interest in collaborating with Indian authorities on affordable AI solutions. The timing of Altman’s praise coincides with a surge in Indian users on OpenAI’s platforms, now the company’s second-largest market.

Meanwhile, OpenAI faces a legal tussle with several Indian media outlets over their alleged content misuse. Despite this, the potential of India’s booming AI market—projected to hit $8 billion by 2025—makes the country a critical frontier for global tech firms.

Experts argue that Altman’s overtures are more about business than sentiment. With increasing competition from rival AI models like DeepSeek and Gemini, maintaining and growing OpenAI’s Indian user base has become vital. As technology analyst Nikhil Pahwa said, ‘There’s no real love; it’s just business.’

For more information on these topics, visit diplomacy.edu.

TikTok deal stalled amid US-China trade tensions

Negotiations to divest TikTok’s US operations have been halted following China’s indication that it would not approve the deal. The development came after President Donald Trump announced increased tariffs on Chinese imports.

The proposed arrangement involved creating a new US-based company to manage TikTok’s American operations, with US investors holding a majority stake and ByteDance retaining less than 20%. This plan had received approvals from existing and new investors, ByteDance, and the US government.

In response to the stalled negotiations, President Trump extended the deadline for ByteDance to sell TikTok’s US assets by 75 days, aiming to allow more time for securing necessary approvals.

He emphasised the desire to continue collaborating with TikTok and China to finalise the deal, expressing a preference to avoid shutting the app in the US.

The future of TikTok in the US remains unpredictable as geopolitical tensions and trade disputes continue to influence the negotiations.

On one side, such a reaction from the Chinese government could have been expected in exchange for the increase of US tariffs on Chinese products; on the other side, by extending the deadline, Trump would be able to maintain his protectionist policy while collecting sympathies from 170 million US citizens using the app, which now is a victim in their eyes as it faces potential banning if the US-China trade war doesn’t calm down and a resolution is not reached within the extended timeframe.

For more information on these topics, visit diplomacy.edu.

Meta unveils Llama 4 models to boost AI across platforms

Meta has launched Llama 4, its latest and most advanced family of open-weight AI models, aiming to enhance the intelligence of Meta AI across services like WhatsApp, Instagram, and Messenger.

Instead of keeping these models cloud-restricted, Meta has made them available for download through its official Llama website and Hugging Face, encouraging wider developer access.

Two models, Llama 4 Scout and Maverick, are now publicly available. Scout, the lighter model with 17 billion active parameters, supports a 10 million-token context window and can run on a single Nvidia H100 GPU.

It outperforms rivals like Google’s Gemma 3 and Mistral 3.1 in benchmark tests. Maverick, the more capable model, uses the same number of active parameters but with 128 experts, offering competitive performance against GPT-4o and DeepSeek v3 while being more efficient.

Meta also revealed the Llama 4 Behemoth model, still in training, which serves as a teacher for the rest of the Llama 4 line. Instead of targeting lightweight use, Behemoth focuses on heavy multimodal tasks with 288 billion active parameters and nearly two trillion in total.

Meta claims it outpaces GPT-4.5, Claude Sonnet 3.7, and Gemini 2.0 Pro in key STEM-related evaluations.

These open-weight AI models allow local deployment instead of relying on cloud APIs, though some licensing limits may apply. With Scout and Maverick already accessible, Meta is gradually integrating Llama 4 capabilities into its messaging and social platforms worldwide.

For more information on these topics, visit diplomacy.edu.

Diplo chatbot can make mistakes. Consider checking important information.