ai

New Jersey criminalises the harmful use of AI deepfakes

New Jersey has become one of several US states to criminalise the creation and distribution of deceptive AI-generated media, commonly known as deepfakes. Governor Phil Murphy signed the legislation on Wednesday, introducing civil and criminal penalties for those who produce or share such media.

If deepfakes are used to commit further crimes like harassment, they may now be treated as a third-degree offence, punishable by fines up to $30,000 or up to five years in prison.

The bill was inspired by a disturbing incident at a New Jersey school where students shared explicit AI-generated images of a classmate.

Governor Murphy had initially vetoed the legislation in March, calling for changes to reduce the risk of constitutional challenges. Lawmakers later amended the bill, which passed with overwhelming support in both chambers.

Instead of ignoring the threat posed by deepfakes, the law aims to deter their misuse while preserving legitimate applications of AI.

‘This legislation takes a proactive approach,’ said Representative Lou Greenwald, one of the bill’s sponsors. ‘We are safeguarding New Jersey residents and offering justice to victims of digital abuse.’

A growing number of US states are taking similar action, particularly around election integrity and online harassment. While 27 states now target AI-generated sexual content, others have introduced measures to limit political deepfakes.

States like Texas and Minnesota have banned deceptive political media outright, while Florida and Wisconsin require clear disclosures. New Jersey’s move reflects a broader push to keep pace with rapidly evolving technology and its impact on public trust and safety.

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

Microsoft rethinks AI data centre strategy amid market shifts

Microsoft has reportedly scaled back or delayed several major data centre projects, just three months after announcing plans to invest $80 billion in AI infrastructure through the current fiscal year.

According to Bloomberg, the company has paused developments in multiple locations, including Australia, Indonesia, the United Kingdom, and US states such as Illinois, North Dakota, and Wisconsin.

Instead of denying the report, Microsoft confirmed adjustments to its plans, citing the need for long-term flexibility. A spokesperson said the company continuously reviews future infrastructure needs to ensure alignment with growing AI demand, adding that the changes reflect Microsoft’s adaptable strategy.

The halted projects include negotiations for high-performance AI chip facilities in the UK and a site near Chicago, along with construction delays in Jakarta and Wisconsin.

These moves come amid growing scrutiny over whether the AI sector is entering a bubble, especially as emerging models challenge the assumption that vast computing power is always necessary for innovation.

Instead of sticking to high-cost development, Microsoft may be responding to a new trend: efficient, lower-cost AI models from Chinese firms that rival those of Western tech giants.

With AI development costs dropping and access expanding, Microsoft’s strategic pause could reflect a shift towards a more sustainable and agile future in AI infrastructure.

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

India among few developing nations with strong AI investment

India and China were the only developing nations to attract notable private investment in AI in 2023, according to the UN’s Technology and Innovation Report 2025. Instead of the US simply leading the field, it dominated with $67 billion in AI investment, accounting for 70 per cent of the global total.

China followed with $7.8 billion, while India ranked tenth worldwide with $1.4 billion. Instead of being evenly distributed, access to AI infrastructure and research remains heavily concentrated in a handful of countries, mainly the US and China.

India’s rise in the AI space stems from policy-driven innovation and education rather than organic growth alone. It climbed to 36th place out of 170 on the UNCTAD Frontier Technologies Readiness Index in 2024, improving from 48th in 2022.

Instead of only focusing on economic size, the index measures readiness through ICT availability, skills, R&D, industrial capacity, and financing. India performed well in R&D and industrial capacity but fell behind in ICT access and skill development.

India has supported its AI ecosystem through collaboration between the government, academia, and the private sector. The country hosts a large developer base, around 13 million, and contributes actively to generative AI projects on platforms like GitHub.

Programmes such as the India AI Mission aim to boost AI education and innovation in smaller cities, instead of keeping progress limited to major urban centres. Institutes like IIT Hyderabad and IIT Kharagpur were named among the country’s key centres of AI excellence.

Still, India faces challenges in expanding its AI capabilities across all sectors. Instead of allowing AI to widen inequalities, the report urges investment in workforce reskilling and inclusion. While AI can boost productivity, it may also displace jobs unless paired with supportive policies.

The technology, if harnessed wisely, could create new industries and strengthen employment rather than replace it.

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

Lumai raises $10 million to revolutionise AI with optical technology

AI accelerator startup Lumai has secured over $10 million in funding to advance its optical computing technology.

The investment round was led by Constructor Capital, with support from IP Group and new investors including PhotonVentures, Journey Ventures, and State Farm Ventures. The funding will be used to expand Lumai’s workforce, develop new products, and establish a stronger presence in the United States market.

Spun out of the University of Oxford, Lumai aims to tackle the growing energy demands of AI by replacing traditional silicon-based computing with optical processing.

Its technology processes AI’s core arithmetic operations using beams of light in 3D space, bypassing the limitations of conventional GPUs.

By using low-cost optical components, Lumai claims its accelerators can cut AI inference costs to a tenth of current solutions while delivering 50 times the performance and consuming just 10% of the power.

The demand for AI computing power is increasing rapidly, with US data centres expected to triple their energy consumption by 2028.

Lumai’s CEO, Tim Weil, believes optical computing is the key to unlocking the future of AI, enabling more sustainable and cost-effective advancements in large language models and next-generation AI.

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

Siemens buys Dotmatics to boost AI drug research

Siemens announced on Wednesday its acquisition of US software firm Dotmatics for $5.1 billion, aiming to enhance its AI capabilities for drug discovery.

The German company described the deal as complementary to its expansion into Life Sciences, positioning itself in a market increasingly reliant on digital transformation to meet growing medical needs.

Siemens expects Dotmatics to generate $100 million annually in the mid-term, rising to $500 million in the long run, and said the acquisition would be immediately profitable. The transaction is set to be completed in the first half of next year.

Founded in 2005, Dotmatics employs 800 people and specialises in AI-driven R&D software designed to accelerate drug research. This move follows Siemens’ recent $10 billion purchase of another AI-powered US software firm, Altair Engineering.

As Siemens’ industrial software faces slowing demand, its digital division has been driving revenue growth instead of its traditional factory automation products. The company, Germany’s second-largest by market value, continues expanding its software portfolio to capitalise on AI-driven innovations.

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

Are digital taxes the new frontline in global trade warfare?

While President Trump’s tariffs on goods dominate headlines, a more consequential battle is brewing over digital services. US tech giants like Meta, Google, and Amazon wield unparalleled global dominance in this sector.

In just-in-time analysis, Jovan Kurbalija argues that Trump’s fixation on traditional trade levers (steel, cars) overlooks a critical vulnerability for the United States: the use of digital services taxes (DSTs) and regulatory pressure by the EU and other trading partners to counterbalance new US tariff.

The collapse of OECD-led multilateral tax negotiations in 2024 has triggered a resurgence of unilateral DSTs, from Canada’s retroactive levy to India’s expanded ‘equalization levy’ and revived EU proposals for bloc-wide digital taxes.

Kurbalija analyses how digital taxation redefines trade diplomacy, with implications ranging from recalibrated leverage (host nations exploiting US tech dependence) to governance gaps (WTO rules ill-equipped for digital disputes). It poses new challenges for digital diplomacy, AI negotiations, and internet governance.

He warns that failure to address this ‘invisible trade war’ could escalate tit-for-tat measures, jeopardizing both physical goods trade and the digital economy. The rise of data and sovereignty will be inevitable.

Ultimately, the piece underscores a paradigm shift: in the 21st-century economy, algorithms, and data flows are as strategically vital as steel beams—and more impactful for economic well-being and global prosperity.

For more information on these topics, read more here.

Anthropic introduces Claude to revolutionise learning and teaching

Claude for Education, launched by Anthropic, introduces a specialised AI for higher education, aiming to support universities in teaching, learning, and administration.

The initiative includes key features like Learning mode, full campus access for top universities, and partnerships with organisations like Internet2 and Instructure to integrate AI into academic tools.

Learning mode helps students develop critical thinking by guiding them through problems with Socratic questioning instead of providing direct answers. It also offers templates for research and study.

Key academic partnerships include Northeastern University, London School of Economics, and Champlain College, all of which will benefit from campus-wide access to Claude. These partnerships ensure AI’s responsible integration and accessibility for all students.

New student programs, such as the Claude Campus Ambassadors and API credit initiatives, provide opportunities for students to engage with and build on AI tools.

The launch also coincides with efforts to integrate AI into the academic plans of institutions like Northeastern University, which is pioneering AI adoption in higher education with its ‘Northeastern 2025’ initiative.

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

Chinese tech firm Honor bets big on AI devices

Chinese smartphone maker Honor has unveiled a $10 billion investment plan aimed at advancing AI technologies across its product ecosystem. The announcement was made by CEO James Li at the Mobile World Congress in Barcelona, where he outlined the company’s ambition to evolve beyond smartphones and expand into AI-powered PCs, tablets, and wearables.

The major funding initiative comes as Honor prepares for a public listing, following a shareholder restructuring completed in December. While a date for the IPO has yet to be confirmed, the company appears to be positioning itself as a key player in China’s AI race, spurred by growing domestic interest in large language models like those developed by DeepSeek.

Despite slipping to fourth place in China’s smartphone market last year, Honor continues to receive strong backing from the Shenzhen local government. Support has included R&D funding, tax incentives, and assistance with international expansion. The company’s strategic pivot to AI reflects broader trends in China’s tech sector, as firms seek to integrate smart features into a wider range of consumer devices.

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

Bill Gates foresees AI-driven two-day work weeks

Bill Gates has predicted that AI will reduce work hours significantly, with two-day work weeks becoming a reality in the next decade.

His vision contrasts with views in India, where figures like Narayana Murthy and S.N. Subrahmanyan argue for longer work hours instead of reducing them, believing they are crucial for economic growth.

However, Harsh Goenka and Harish Mariwala, who focus on ambition and productivity instead of hours worked, have a different take. They stress the importance of quality instead of quantity in work hours.

Studies, such as those from Iceland and New Zealand, suggest that reduced working hours can lead to higher productivity and a better work-life balance.

In Iceland, trials reducing weekly hours from 40 to 35 saw workers feeling more energised and less stressed, while New Zealand’s Perpetual Guardian found that employees completed tasks in fewer hours, raising engagement and reducing stress.

Despite Gates’ prediction, the idea of working only two days a week seems extreme, especially considering his past work ethic.

Gates himself worked long hours instead of fewer ones during his early career, fearing mistakes could cost him his company, but he acknowledges that AI and technology have shifted the landscape.

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

The overlooked legacy of the Lwów–Warsaw school in shaping modern AI

AI may seem like a modern marvel, but its foundations stretch back to early 20th-century philosophical thought—particularly to the Lwów–Warsaw School of Philosophy. In a recent blog by Jovan Kurbalija, Executive Director of Diplo, he highlights how this influential movement, established by Kazimierz Twardowski and active between the world wars, made groundbreaking contributions to logic, semantics, and analytical philosophy—fields that underpin today’s AI systems.

Despite being overshadowed by more famous contemporaries like the Vienna Circle, the Lwów–Warsaw School cultivated thinkers whose work continues to shape how machines process language and make decisions. Among the most influential figures was Alfred Tarski, whose formal definition of truth laid the groundwork for natural language processing (NLP), enabling AI to understand and generate human language.

Another key contributor, Jan Łukasiewicz, developed many-valued logic—a system that goes beyond simple true/false decisions and allows machines to reason through uncertainty, essential in complex tasks like autonomous driving or medical diagnostics. The school’s importance isn’t just technical.

Its strong emphasis on analytical rigour also offers tools for navigating AI’s growing ethical concerns—ranging from fairness in algorithmic decisions to broader questions about consciousness and societal impact. Though these philosophical aspects are less often discussed, they’re crucial for ensuring AI development remains responsible and aligned with human values.

The intellectual legacy is rarely acknowledged in mainstream AI history. Yet, as Kurbalija emphasises, AI’s roots are deeper and more diverse than we often realise. Revisiting the Lwów–Warsaw School highlights how past ideas can still guide us in shaping the technologies of tomorrow.

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