A group of researchers from Hao AI Lab at the University of California San Diego has suggested that Super Mario Bros. might actually be a tougher challenge for AI than Pokémon. In a recent experiment, AI models were tasked with playing the game, and while Anthropic’s Claude 3.7 performed the best, models like Google’s Gemini 1.5 Pro and OpenAI’s GPT-4o struggled. The game was not the original 1985 version but instead ran in an emulator integrated with GamingAgent, a framework that provided basic instructions and screenshots for the AI to control Mario.
The AI had to generate inputs, such as Python code, based on the given instructions to navigate Mario through the game’s challenges. The researchers found that while the game required models to plan complex manoeuvres and strategies, reasoning models like OpenAI’s o1 performed worse than non-reasoning models. This is because reasoning models typically take longer to decide on actions, and in a real-time game like Super Mario Bros., timing is critical.
While games have long been used to benchmark AI, some experts question the relevance of gaming skills as a measure of technological advancement. Andrej Karpathy, a research scientist at OpenAI, has expressed concerns over the current AI evaluation process, calling it an ‘evaluation crisis.’ Despite these concerns, watching AI take on Super Mario Bros. provides an interesting glimpse into how far AI has come, even if the benchmarks remain unclear.
For more information on these topics, visit diplomacy.edu.
Amazon has announced plans to use AI to help reduce flood risks in Spain’s northeastern region of Aragon, where it is building new data centres.
As part of its $17.9 million investment, Amazon’s cloud computing unit AWS will modernise infrastructure and optimise agricultural water use to tackle flood concerns.
The move follows catastrophic floods that impacted large areas around Valencia and comes as AWS continues its €15.7 billion expansion in the region’s cloud infrastructure.
The region is prone to flooding, especially along the Ebro River, highest-flow river in Spain, which crosses through Aragon on its way to the Mediterranean.
Amazon will deploy advanced cloud computing technologies to create an early warning system combining real-time data collection, sensor networks, and AI-powered analysis.
However, this system will help Zaragoza, the capital of Aragon, monitor flood risks more effectively and provide timely warnings to emergency services.
In addition to its technological investment, local authorities in Zaragoza are building flood defences at the Barranco de la Muerte, or Death Ravine, to mitigate future flood damage.
With these combined efforts, Amazon aims to contribute to reducing the region’s vulnerability to floods while supporting its own expanding data infrastructure.
For more information on these topics, visit diplomacy.edu.
Chinese AI startup Zhipu AI has raised over 1 billion yuan ($137.22 million) in new funding, following a previous 3 billion yuan investment in December. The latest funding round comes amid growing competition in China’s AI sector, particularly with rival DeepSeek’s large language models, which claim to rival Western models at a lower cost.
Investors in the new round include the state-backed Hangzhou City Investment Group Industrial Fund and Shangcheng Capital. Zhipu AI plans to use the funds to enhance its GLM language model and expand its AI ecosystem, with a focus on businesses in Zhejiang province and the Yangtze River Delta region.
Founded in 2019, Zhipu AI has completed 16 funding rounds and is considered one of China’s leading AI startups. The company aims to release a range of new AI models, including foundation and multimodal models, as part of its open-source strategy. The funding comes as competitors like DeepSeek continue to disrupt the AI landscape with open-source models that challenge established platforms.
For more information on these topics, visit diplomacy.edu.
Vodafone UK has teamed up with IBM to explore quantum-safe cryptography as part of a new Proof of Concept (PoC) test for its mobile and broadband services, particularly for users of its ‘Secure Net’ anti-malware service. While quantum computers are still in the early stages of development, they could eventually break current internet encryption methods. In anticipation of this, Vodafone and IBM are testing how to integrate new post-quantum cryptographic standards into Vodafone’s existing Secure Net service, which already protects millions of users from threats like phishing and malware.
IBM’s cryptography experts have co-developed two algorithms now recognised in the US National Institute of Standards and Technology’s first post-quantum cryptography standards. This collaboration, supported by Akamai Technologies, aims to make Vodafone’s services more resilient against future quantum computing risks. Vodafone’s Head of R&D, Luke Ibbetson, stressed the importance of future-proofing digital security to ensure customers can continue enjoying safe internet experiences.
Although the PoC is still in its feasibility phase, Vodafone hopes to implement quantum-safe cryptography across its networks and products soon, ensuring stronger protection for both business and consumer users.
For more information on these topics, visit diplomacy.edu.
British startup Wayve has announced plans to open a new testing and development hub in Germany, deploying a fleet of test vehicles in the Stuttgart region. The self-driving technology firm aims to enhance features like lane change assistance at the new facility, which will focus on improving its “Embodied AI” system that learns from human behaviour.
Wayve, which operates in the UK and the US, is expanding into Germany as part of its strategy to enter the European market, particularly Germany, the continent’s largest automotive hub. The company received a boost earlier this year, with Uber investing in August and SoftBank leading a $1 billion funding round in May, supported by Nvidia.
Despite the significant investments in autonomous vehicle technology, self-driving systems still face challenges in predicting and assessing risks as accurately as human drivers. Wayve’s technology is already integrated into six vehicle platforms, including electric models like the Jaguar I-PACE and Ford Mustang Mach-E, as part of advanced driver assistance systems (ADAS).
For more information on these topics, visit diplomacy.edu.
Singapore‘s government is investigating a fraud case involving servers supplied by US companies, potentially containing Nvidia’s advanced chips. Three men, including a Chinese national, were charged last week in connection with the alleged illegal transfer of these AI chips from Singapore to Chinese firm DeepSeek. Singapore’s Home Affairs and Law Minister K Shanmugam confirmed that the servers were provided by Dell Technologies and Super Micro Computer to local firms before being sent to Malaysia.
Authorities are still unsure whether Malaysia was the final destination, but they are working with US officials to determine if the servers contained restricted US export-controlled items. The US is already investigating whether DeepSeek has used banned Nvidia chips, which could lead to violations of export laws. The case forms part of a broader probe into suspected smuggling activities linked to AI chips being moved from countries like Singapore to China.
Singapore, a key market for Nvidia, is also examining allegations that DeepSeek may have acquired thousands of advanced Nvidia chips illegally. However, DeepSeek has denied these claims, stating that it only used legally purchased chips, including the Nvidia H800 model. The investigation is ongoing, with Singapore continuing to cooperate with US authorities.
For more information on these topics, visit diplomacy.edu.
Chinese smartphone maker Honor plans to invest $10 billion over the next five years to develop AI for its devices as it prepares for a public listing. CEO James Li revealed at the Mobile World Congress in Barcelona that the company aims to expand beyond smartphones into AI-powered PCs, tablets, and wearables.
Honor’s push into AI comes after completing a shareholder restructuring in December, bringing it closer to an initial public offering, though no timeline has been set. The announcement also aligns with a surge in AI investment in China, driven by the popularity of DeepSeek’s affordable language models.
While Honor’s market share in China slipped from second to fourth place last year, it has gained significant support from the Shenzhen government, including R&D funding and tax breaks. The company is also focusing on expanding internationally as it strengthens its AI capabilities.
For more information on these topics, visit diplomacy.edu.
Britain’s privacy regulator, the Information Commissioner’s Office (ICO), has launched an investigation into the child privacy practices of TikTok, Reddit, and Imgur. The ICO is scrutinising how these platforms manage personal data and age verification for users, particularly teenagers, to ensure they comply with UK data protection laws.
The investigation focuses on TikTok’s use of data from 13-17-year-olds to recommend content via its algorithm. The ICO is also examining how Reddit and Imgur assess and protect the privacy of child users. If evidence of legal breaches is found, the ICO will take action, as it did in 2023 when TikTok was fined £12.7 million for mishandling data from children under 13.
Both Reddit and Imgur have expressed a commitment to adhering to UK regulations. Reddit, for example, stated that it plans to roll out updates to meet new age-assurance requirements. Meanwhile, TikTok and Imgur have not yet responded to requests for comment.
The investigation comes amid stricter UK legislation aimed at safeguarding children online, including measures requiring social media platforms to limit harmful content and enforce age checks to prevent underage access to inappropriate material.
For more information on these topics, visit diplomacy.edu.
Chinese technology startups are scrambling for fresh funding to capitalise on the growing excitement around AI, fueled by DeepSeek‘s AI breakthrough and President Xi Jinping’s recent support for private enterprises. Companies such as AI-powered optics startup Rid Vision, brain-computer interface firm AI CARE Medical, and robotics maker Shanghai Qingbao Engine Robotics are among those seeking new onshore investments, according to venture capitalist Andrew Qian.
The surge in investment interest comes after a long period of stagnation in China‘s venture capital sector. New Access Capital, which has invested in several AI startups, noted that many businesses are now attracting both cooperation discussions and investment offers. The DeepSeek case has highlighted the rise of innovative Chinese tech firms, moving away from previous trends of copycat startups, and offering hope for a revitalised venture capital environment.
Despite the growing investor enthusiasm, challenges remain. The strict regulatory scrutiny of initial public offerings (IPOs) in China, coupled with heightened tensions between China and the US, complicate the prospects for exit strategies. Venture capital firms are particularly focused on AI-related businesses, with several companies in sectors like AI image generation and medical tech securing significant funding in recent weeks.
However, the overall fundraising environment remains difficult. Data shows a sharp decline in venture capital investments since 2021, and while investor sentiment has improved post-DeepSeek, concerns over the regulatory landscape and geopolitical tensions persist. Some analysts remain cautious about the near-term outlook for IPOs, especially for offshore listings.
For more information on these topics, visit diplomacy.edu.
Three men have been charged with fraud in Singapore as part of an investigation that may be linked to the illegal transfer of advanced Nvidia chips to the Chinese AI company, DeepSeek. The United States is also probing whether DeepSeek used US-made chips, which are prohibited from being shipped to China. This case is part of a wider investigation into potential AI chip smuggling operations involving several individuals and companies across countries, including Singapore.
The charges focus on two Singaporeans, Aaron Woon Guo Jie and Alan Wei Zhaolun, who are accused of falsely claiming that Nvidia chips would not be transferred to unauthorised recipients. A third man, Chinese national Li Ming, is charged with committing similar fraud regarding the shipment of servers. These actions are believed to be connected to DeepSeek’s use of Nvidia chips, although the authorities have not confirmed the company’s involvement.
If convicted, the men could face up to 20 years in prison, a fine, or both. Singapore’s authorities have been actively investigating the smuggling network, with police recently arresting nine individuals in a raid across 22 locations, seizing records and electronic evidence.
Nvidia’s business in Singapore is substantial, making it the company’s second-largest market after the US. However, only a small percentage of its revenue comes from direct shipments to the country. Singapore’s government has emphasised its commitment to strict enforcement of export controls, underscoring that the nation will not tolerate any attempts to evade international regulations.
For more information on these topics, visit diplomacy.edu.
