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.
A court in India has ordered an Amazon unit to pay $39 million in damages for trademark infringement after unauthorised ‘Beverly Hills Polo Club’ garments were sold on its platform.
The ruling marks one of the highest damages awarded against a US company in an Indian trademark case. Lifestyle Equities, the owner of the Beverly Hills Polo Club brand, filed the lawsuit in 2020, alleging that Amazon’s Indian website was selling apparel with nearly identical branding at lower prices.
The Delhi High Court found that the infringing brand was owned by Amazon Technologies and sold directly on the platform. Judges noted that the logos were ‘hardly distinguishable’ and stated that Amazon was aware of the brand’s exclusive rights, citing previous legal disputes in the United Kingdom.
The ruling includes a permanent injunction, barring Amazon from selling counterfeit products with the Beverly Hills Polo Club branding. The company has denied wrongdoing, and its representatives have not commented on the decision.
Amazon has faced similar legal battles before, including a trademark dispute in London, where it lost an appeal last year. Critics have long accused the e-commerce giant of engaging in predatory practices, with a 2021 Reuters investigation alleging that Amazon manipulated search results to promote its own private brands in India.
The recent court decision has reignited calls for stricter enforcement of trademark laws and fair business practices in the country.
For more information on these topics, visit diplomacy.edu.
The University of Cambridge is set to open a new department, The Bennett School of Public Policy, later this year, dedicated to exploring how AI can improve both business and public sector operations.
However, this marks the university’s first major new academic department in the 21st century, with research priorities focused on harnessing AI to address pressing policy challenges.
Professors from a wide range of disciplines, including economics, social sciences, and medical fields, will collaborate to investigate the potential benefits and risks of AI adoption across sectors.
The school’s leadership aims to foster a generation of ‘tech-savvy’ policymakers who can effectively manage AI’s role in society while promoting sustainable and inclusive growth.
In addition to investigating AI’s economic impact, researchers will work closely with the Civil Service to develop AI workflows and help local government leaders address geographical inequality.
With a focus on cross-disciplinary expertise, the Bennett School plans to train policymakers who can tackle the complex issues of today’s world.
For more information on these topics, visit diplomacy.edu.
Lenovo has won an appeal in a UK court that will allow it to secure a temporary licence for Ericsson’s patents, marking a significant development in the ongoing patent dispute between the two companies.
The case, which revolves around fair, reasonable, and non-discriminatory (FRAND) licensing terms for 4G and 5G wireless technology, has seen both companies take legal action in various countries, including the UK, Brazil, and the US.
In his ruling, Judge Richard Arnold determined that Ericsson had failed to act in good faith by pursuing legal claims in foreign courts despite Lenovo’s willingness to accept the FRAND terms set by the English courts.
He stated that, as a willing licensor, Ericsson should have agreed to an interim licence, with Lenovo being required to pay a substantial sum to Ericsson. Lenovo’s Chief Legal Officer hailed the decision as a victory for transparency and fairness in global patent licensing.
The ruling follows Lenovo’s 2023 lawsuit against Ericsson in the UK, a part of the broader dispute between the two over the terms for the use of each other’s patents. Ericsson has yet to comment on the decision.
For more information on these topics, visit diplomacy.edu.
Twenty-three industry groups across Europe have urged EU officials to approve a draft cybersecurity certification for cloud services.
The certification scheme, which was introduced in 2020 by the European Union Agency for Cybersecurity (ENISA), aims to provide governments and businesses with reliable, secure cloud service providers.
It has been under revision since last year, with changes that favour major tech firms such as Microsoft, Google, and Amazon.
The groups, including the American Chamber of Commerce and various EU trade associations, argue that the updated draft, which will be finalised in March 2024, strikes a balance between robust security standards and an open market.
These revisions removed political provisions, such as requirements for US tech companies to partner with EU firms for data storage, focusing instead on technical security criteria.
Despite this, concerns linger about the European Commission potentially delaying or even scrapping the certification scheme altogether. The groups have strongly urged the EU to push forward with its adoption, believing it will support Europe’s digital economy while promoting secure cloud computing solutions.
For more information on these topics, visit diplomacy.edu.
The notice relates to the acquisition of two companies, Little Internet Private Limited and Nearbuy India Private Limited, covering the period from 2015 to 2019.
Paytm has stated that the alleged violations occurred before these companies became its subsidiaries. The company maintains that the notice has no impact on its services for consumers and merchants. No further details about the specific allegations have been disclosed.
The notice comes at a time of increased regulatory scrutiny of digital payment firms in India.
Paytm has previously faced challenges related to compliance and operations, making the latest development a significant issue for the company as it continues to navigate the country’s evolving financial regulations.
For more information on these topics, visit diplomacy.edu.
The founder of cryptocurrency financial services firm Gotbit has been extradited from Portugal to the United States to face charges of market manipulation and fraud.
Aleksei Andriunin, a 26-year-old Russian national, appeared in a Boston court, where he pleaded not guilty to wire fraud and conspiracy charges. Prosecutors allege his company engaged in sham trading to artificially inflate the value of digital tokens.
Authorities claim that between 2018 and 2024, Gotbit manipulated cryptocurrency trading volumes through a practice known as ‘wash trading’. The FBI’s ‘Operation Token Mirrors‘ played a key role in the investigation, reportedly using its own digital token to detect fraudulent activities.
Gotbit allegedly facilitated wash trades worth millions and profited tens of millions of dollars by boosting the market appeal of cryptocurrencies such as Saitama and Robo Inu.
Andriunin was arrested in Portugal in October when US authorities first announced charges against him and others. His company and two employees in Russia also face legal action, though they have not yet appeared in court.
The case is part of a broader crackdown on fraud in the cryptocurrency sector.
For more information on these topics, visit diplomacy.edu.
United States officials are reviewing whether the UK breached a bilateral agreement by reportedly pressuring Apple to create a ‘backdoor’ for government access to encrypted iCloud backups.
Apple recently withdrew an encrypted storage feature for UK users following reports that it had refused to comply with such demands, which could have affected users worldwide. The Washington Post reported that Apple rejected the UK government’s request.
The US director of national intelligence, Tulsi Gabbard, confirmed in a letter to lawmakers that a legal review is underway to determine if the UK violated the CLOUD Act.
Under the agreement, neither the US nor the United Kingdom can demand data access for citizens or residents of the other country. Initial legal assessments suggest the UK’s reported demands may have overstepped its authority under the agreement.
Apple has long defended its encryption policies, arguing that creating a backdoor for government access would weaken security and leave user data vulnerable to hackers. Cybersecurity experts warn that any such backdoor, once created, would inevitably be exploited.
The tech giant has clashed with regulators over encryption before, notably in 2016 when it resisted US government efforts to unlock a terrorism suspect’s iPhone.
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.
