ai

Politeness to AI is about us, not them

In his thought-provoking blog post ‘Politeness in 2025: Why are we so kind to AI?’, Dr Jovan Kurbalija explores why nearly 80% of users in the UK and the USA instinctively say ‘please’ and ‘thank you’ to AI platforms like ChatGPT.

While machines lack feelings, our politeness reveals more about human psychology and cultural habits than the technology itself. For many, courtesy is a deeply ingrained reflex shaped by personality traits such as agreeableness and lifelong social conditioning, extending kindness even to non-sentient entities.

However, not everyone shares this approach. Some users are driven by subtle fears of future AI dominance, using politeness as a safeguard, while others prioritise efficiency, viewing AI purely as a tool undeserving of social niceties.

A rational minority dismisses politeness altogether, recognising AI as nothing more than code. Dr Kurbalija highlights that these varied responses reflect how we perceive and interact with technology, influenced by both evolutionary instincts and modern cognitive biases.

Beyond individual behaviour, Kurbalija points to a deeper issue: our tendency to humanise AI and expect it to behave like us, unlike traditional machines. This blurring of lines between tool and teammate raises important questions about how our perceptions shape AI’s role in society.

Ultimately, he suggests that politeness toward AI isn’t about the machine—it reflects the kind of humans we aspire to be, preserving empathy and grace in an increasingly digital world.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Google boosts AI in coding and cloud growth

More than 30% of all code at Google is now written with the help of AI, according to CEO Sundar Pichai during Alphabet’s Q1 2025 earnings call.

Developers are embracing AI-generated suggestions in nearly one out of every three code changes, thanks to improved models and the rollout of agentic workflows—AI systems designed to manage complex, multi-step tasks.

Pichai noted that AI-assisted coding is expanding across the company, with customer service teams leading the way in adoption.

Alphabet reported strong financial results, with quarterly revenue climbing 12% year-over-year to $90.2 billion. Net income rose sharply by 46% to $34.5 billion, and earnings per share jumped 49%.

While there was a slight quarterly dip in revenue from the previous quarter, Google’s core advertising business, YouTube, and Google Cloud all contributed to year-on-year growth. Cloud revenue, in particular, increased 28% due to surging demand for AI and infrastructure products.

The recently released Gemini 2.5 Pro model was described by Pichai as ‘state-of-the-art’, outperforming rivals across benchmarks and landing the top spot on Chatbot Arena. Gemini models are now integrated across 15 Google products used by more than half a billion people.

Features like Gemini Live and AI-powered camera tools are being rolled out on Android and Pixel devices, while Google Assistant will also receive a Gemini upgrade later this year.

Instead of slowing down, Google is accelerating its AI development with initiatives like Gemini Robotics Models, the AI Co-Scientist for research, and the continued success of AlphaFold, used by over 2.5 million scientists.

With over 200% growth in users of AI Studio and Gemini API, and more than 140 million downloads of Gemma models, Google is clearly positioning AI at the centre of its future across products, platforms, and research.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

ChatGPT expands Deep Research to more users

A new feature introduced by ChatGPT in February, called Deep Research, is gradually becoming available across its user base. This includes subscribers on the Plus, Team, and Pro plans, while even those using the free ChatGPT app on iOS and Android can now access a simplified version.

Designed to carry out in-depth reports and analyses within minutes, Deep Research uses OpenAI’s o3 model to perform tasks that would otherwise take people hours to complete.

Instead of limiting access to paid users alone, OpenAI has rolled out a lightweight version powered by its o4-mini AI model for free users. Although responses are shorter, the company insists the quality and depth remain comparable.

The more efficient model also helps reduce costs, while delivering what OpenAI calls ‘nearly as intelligent’ results as the full version.

The feature’s capabilities stretch from suggesting personalised product purchases like cars or TVs, to helping with complex decisions such as choosing a university or analysing market trends.

Free-tier users are currently allowed up to five Deep Research tasks each month, whereas Plus and Team plans get ten full and fifteen lightweight tasks. Pro users enjoy a generous 125 tasks of each version per month, and EDU and Enterprise plans will begin access next week.

Once users hit their full version limit, they’ll be automatically shifted to the lightweight tool instead of losing access altogether. Meanwhile, Google’s GeminiAI offers a similar function for its paying customers, also aiming to deliver quick, human-level research and analysis.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Google stopped Motorola from using Perplexity as default assistant

A senior executive at Perplexity AI has testified that Google prevented Motorola from setting the AI startup’s assistant as the default on new smartphones, citing restrictive contracts.

The testimony came during a trial seeking remedies after Google was found to have unlawfully maintained a search monopoly.

Although Motorola will preload the app, it cannot make it the default due to binding agreements with Google. Perplexity’s Chief Business Officer, described the difficulty of replacing Google’s assistant on Android phones, saying Google’s terms create an environment where device makers fear losing revenue.

The CEO added that ongoing negotiations with other companies only became possible due to pressure from the US Department of Justice’s antitrust case.

The Justice Department is asking the court to ban Google from paying for default placements, which would also affect its AI products like Gemini.

Meanwhile, Perplexity is developing its own browser, Comet, and voiced concern about any Chrome sale undermining open-source access. The company does not support OpenAI’s interest in acquiring the browser, citing past inconsistencies in its open-source commitments.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Legal AI startup Noxtua backed by top law firms

Germany-based legal tech firm Noxtua SE has secured €80.7 million in funding to support the launch of its AI-powered legal workspace, Beck-Noxtua. Previously known as Xayn SE, the company announced the round following a major rebrand.

C.H. Beck, Germany’s leading legal publisher, led the investment, with backing from Northern Data AG, CMS LTF Limited, Dentons Corp., and other strategic partners.

Beck-Noxtua is designed to streamline legal workflows by offering a suite of tools powered by large language models trained on over 55 million documents from C.H. Beck’s legal library.

These include the Noxtua Legal LLM, capable of generating and analysing complex legal texts, and Noxtua Voyage Embed, which enhances legal search and precedent discovery. The platform complies with the EU’s General Data Protection Regulation and is aimed at law firms across Europe.

Noxtua plans to expand beyond Germany, establishing offices in other EU countries. With support from Northern Data’s European infrastructure and legal input from firms like CMS and Dentons, the company is positioning itself as a key player in the continent’s legal AI space.

Analysts note Noxtua’s strategic partnerships as a potential advantage, though caution that the legal industry’s resistance to technology may prove a challenge.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Microsoft expands rewards for reporting AI vulnerabilities

Microsoft has announced an expanded bug bounty initiative, offering up to $30,000 for researchers who uncover critical vulnerabilities in AI features within Dynamics 365 and the Power Platform.

The programme aims to strengthen security in enterprise software by encouraging ethical hackers to identify and report risks before cybercriminals can exploit them.

Rather than relying on general severity scales, Microsoft has introduced an AI-specific vulnerability classification system. It highlights prompt injection attacks, data poisoning during training, and techniques like model stealing and training data reconstruction that could expose sensitive information.

Highest payouts are reserved for flaws that allow attackers to access other users’ data or perform privileged actions without their consent.

The company urges researchers to use free trials of its services, such as PowerApps and AI Builder, to identify weaknesses. Detailed product documentation is provided to help participants understand the systems they are testing.

Even reports that don’t qualify for a financial reward can still lead to recognition if they result in improved defences.

The AI bounty initiative is part of Microsoft’s wider commitment to collaborative cybersecurity. With AI becoming more deeply integrated into enterprise software, the company says it is more important than ever to identify vulnerabilities early instead of waiting for security breaches to occur.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Ransomware decline masks growing threat

A recent drop in reported ransomware attacks might seem encouraging, yet experts warn this is likely misleading. Figures from the NCC Group show a 32% decline in March 2025 compared to the previous month, totalling 600 incidents.

However, this dip is attributed to unusually large-scale attacks in earlier months, rather than an actual reduction in cybercrime. In fact, incidents were up 46% compared with March last year, highlighting the continued escalation in threat activity.

Rather than fading, ransomware groups are becoming more sophisticated. Babuk 2.0 emerged as the most active group in March, though doubts surround its legitimacy. Security researchers believe it may be recycling leaked data from previous breaches, aiming to trick victims instead of launching new attacks.

A tactic like this mirrors behaviours seen after law enforcement disrupted other major ransomware networks, such as LockBit in 2024.

Industrials were the hardest hit, followed by consumer-focused sectors, while North America bore the brunt of geographic targeting.

With nearly half of all recorded attacks occurring in the region, analysts expect North America, especially Canada, to remain a prime target amid rising political tensions and cyber vulnerability.

Meanwhile, cybercriminals are turning to malvertising, malicious code hidden in online advertisements, as a stealthier route of attack. This tactic has gained traction through the misuse of trusted platforms like GitHub and Dropbox, and is increasingly being enhanced with generative AI tools.

Instead of relying solely on technical expertise, attackers now use AI to craft more convincing and complex threats. As these strategies grow more advanced, experts urge organisations to stay alert and prioritise threat intelligence and collaboration to navigate this volatile cyber landscape.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Google launches Gemini 2.5 Flash with faster and smarter AI thinking

Google has unveiled Gemini 2.5 Flash, an enhanced version of its AI model that brings improved performance, faster response times, and cost-efficiency for developers and users alike.

Released in preview through Google AI Studio and Vertex AI, the model is designed to be highly responsive while also offering advanced reasoning capabilities when needed.

The hybrid design allows developers to toggle the model’s ‘thinking’ function, enabling it to pause before answering to better understand prompts and plan complex responses.

Even with this added step, Gemini 2.5 Flash maintains the quick speeds of its predecessor, Gemini 2.0 Flash, while improving on accuracy and depth of output in areas like mathematics, science, coding, and visual reasoning.

Benchmark tests reveal Gemini 2.5 Flash outperforms competing AI models across a range of technical and analytical tasks.

Developers can begin building with it today, and everyday users can explore the model through the Gemini app, which also features earlier models including Gemini 2.0 Flash Thinking and Gemini 2.0 Pro Experimental.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Legal AI firm Noxtua backed by top law publishers

Formerly known as Xayn, the AI startup Noxtua has raised $92.2 million in a Series B round to expand its sovereign AI platform for the legal sector.

The company, which began as a privacy-focused mobile AI app, now specialises in providing legally compliant AI tools for law firms and legal departments in Germany.

The funding round was led by C.H. Beck, Germany’s top legal publisher, which also supplies Noxtua with access to its vast legal database of 55 million documents.

Other investors include Northern Data Group, CMS, and Dentons. Noxtua’s legal AI tool, Beck-Noxtua, will be hosted exclusively on German infrastructure provided by Northern Data to ensure data sovereignty and compliance.

Built from research at Oxford and Imperial College London, Noxtua’s technology is tailored to the complex legal frameworks of Germany and continental Europe.

CEO Dr Leif-Nissen Lundbæk emphasised that American-trained AI models often fail to meet the precision and compliance standards required in Germany, making localised AI solutions vital.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Jamaica tests AI tools to aid teachers

The Jamaican Ministry of Education is testing AI tools in schools to assist teachers with marking and administrative duties.

Portfolio Minister Senator Dana Morris Dixon announced this during the Jamaica Teachers’ Association (JTA) Education Conference 2025, emphasising that AI would allow teachers to focus more on interacting with students, while AI handles routine tasks like grading.

The Ministry is also preparing to launch the Jamaica Learning Assistant, an AI-powered tool that personalises learning to fit individual students’ preferences, such as stories, humour, or quizzes.

Morris Dixon highlighted that AI is not meant to replace teachers, but to support them in delivering more effective lessons. The technology will allow students to review lessons, explore topics in more depth, and reinforce their understanding outside the classroom.

Looking ahead, the Government plans to open Jamaica’s first state-of-the-art AI lab later this year. The facility will offer a space where both students and teachers can develop technological solutions tailored for schools.

Additionally, the Ministry is distributing over 15,000 laptops, 600 smart boards, and 25,000 vouchers for teachers to subsidise the purchase of personal laptops to further integrate technology into the education system.

JTA President Mark Smith acknowledged the transformative potential of AI, calling it one of the most significant technological breakthroughs in history.

He urged educators to embrace this new paradigm and collaborate with the Ministry and the private sector to advance digital learning initiatives across the island.

The conference, held under the theme ‘Innovations in Education Technology: The Imperative of Change,’ reflects the ongoing push towards modernising education in Jamaica.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!