Apple is facing mounting criticism over its AI-generated news summaries, which have produced inaccurate and misleading alerts on its latest iPhones. Media organisations, including the BBC, have raised concerns that the feature, designed to summarise breaking news notifications, has fabricated details that contradict original reports. The National Union of Journalists and Reporters Without Borders have called for the product’s removal, warning it risks spreading misinformation at a time when trust in news is already fragile.
High-profile errors have fuelled demands for urgent action. In December, an Apple AI summary falsely claimed that a murder suspect had taken his own life, while another inaccurately announced Luke Littler as the winner of the PDC World Darts Championship before the event had even begun. Apple has pledged to update the feature to make it clearer that summaries are AI-generated, but critics argue this does not address the root problem.
Journalism watchdogs and industry experts have warned that AI-driven news aggregation remains unreliable. The BBC stressed that the errors could undermine public trust, while former Guardian editor Alan Rusbridger described Apple’s technology as “out of control”. Similar concerns have been raised over generative AI tools from other tech firms, with Google’s AI-powered search summaries also facing scrutiny for producing incorrect responses. Apple insists the feature remains optional and is still in beta testing, with further improvements expected in an upcoming software update.
Recent reports reveal that Chinese hackers have compromised a broader range of US telecommunications companies than previously known. In addition to earlier breaches involving AT&T and Verizon, the cyberattacks have now been found to affect Charter Communications, Consolidated Communications, Windstream, Lumen Technologies, and T-Mobile. The hacking group, identified as Salt Typhoon and linked to Chinese intelligence, exploited vulnerabilities in network devices from security vendors such as Fortinet and Cisco Systems.
The Wall Street Journal reports that US National Security Adviser Jake Sullivan informed telecommunications and technology executives in a confidential meeting in late 2023 that these hackers had developed the capability to disrupt critical US infrastructure, including ports and power grids. While companies like AT&T and Verizon have stated that their networks are now secure and that they are collaborating with law enforcement, concerns persist about the extent and impact of these breaches.
China has denied involvement in these cyber activities, accusing the United States of disseminating disinformation. Nonetheless, the revelations have intensified discussions about national security and the resilience of US critical infrastructure against sophisticated cyber threats. The situation underscores the ongoing challenges in safeguarding sensitive communications and infrastructure from state-sponsored cyber espionage.
Demis Hassabis, CEO of Google DeepMind, has warned that the rapid progress in AI development may be slowing as companies exhaust the available digital data needed to train large language models. The industry has long relied on feeding vast amounts of online text into AI systems to improve performance, but diminishing returns are now setting in. Some experts, including OpenAI’s Ilya Sutskever, believe the industry has reached “peak data,” meaning future improvements will require entirely new approaches.
Researchers are now exploring alternative methods, such as synthetic data, where AI models generate and learn from their own outputs. While this technique has shown promise in fields like mathematics and programming, it struggles with more complex areas like philosophy and the arts, where defining correctness is difficult. OpenAI has already applied this method in its latest system, OpenAI o1, but challenges remain, particularly in preventing AI from making errors or generating misleading information.
Another possibility to overcome ‘data limitation’ in AI development is to shift focus from quantity to quality of data through better data labelling and contextual enrichment, as done by Diplo’s cognitive proximity approach (see below).
DiploAI approach to data enrichment
Tech leaders remain divided on whether AI advancements will continue at the same pace. Nvidia’s CEO Jensen Huang remains optimistic, citing strong demand for AI chips and ongoing innovation. However, some of the company’s biggest customers are preparing for a possible plateau in AI development. Despite the uncertainty, investment in AI infrastructure remains high, with firms continuing to push the boundaries of what AI can achieve.
OpenAI CEO Sam Altman has revealed that the company is losing money on its $200-per-month ChatGPT Pro plan due to unexpectedly high usage. The plan, introduced last year, provides access to an advanced AI model and fewer restrictions on OpenAI’s tools. Altman admitted that the pricing was not based on a rigorous study but was instead a personal decision.
Despite raising around $20 billion, OpenAI remains unprofitable, with estimated losses of $5 billion last year. The company is considering price increases or usage-based fees to improve financial stability. Altman also acknowledged that OpenAI requires more investment than initially expected.
The company remains optimistic about its future revenue, projecting $11.6 billion in 2025 and aiming for $100 billion by 2029. As OpenAI undergoes corporate restructuring, attracting new investors and refining its pricing strategy will be key to long-term profitability.
Avataar, a startup backed by Peak XV and Tiger Global, has launched Velocity, an AI-powered tool that generates product videos directly from a product link. The company aims to help brands create engaging promotional content at scale, competing with tech giants like Amazon and Google, who are also exploring AI-driven video solutions. By automating video production, Avataar hopes to lower costs and improve conversion rates for e-commerce businesses.
Founded in 2015, Avataar initially focused on AI-powered imagery and 3D modelling before shifting towards AI video creation. Its Velocity tool, commercially tested since September 2024, is already being used by brands such as HP, Victoria’s Secret, and Newegg. The company also offers an API, allowing businesses to integrate video generation into their platforms.
Avataar’s models are designed to ensure high-quality product representation, addressing past concerns about AI-generated content accuracy. Investor Amir Konigsberg highlighted the challenge of making AI-generated videos unique and maintaining quality at scale. With advancements in AI models, Avataar believes automated video production is now viable for mass adoption in e-commerce.
OpenAI CEO Sam Altman has stated that the company believes it knows how to build AGI and is now turning its focus towards developing superintelligence. He argues that advanced AI could significantly boost scientific discovery and economic growth. While AGI is often defined as AI that outperforms humans in most tasks, OpenAI and Microsoft also use a financial benchmark—$100 billion in profits—as a key measure.
Despite Altman’s optimism, today’s AI systems still struggle with accuracy and reliability. OpenAI has previously acknowledged that transitioning to a world with superintelligence is far from certain, and controlling such systems remains an unsolved challenge. The company has, however, recently disbanded key safety teams, leading to concerns about its priorities as it seeks further investment.
Altman remains confident that AI will soon make a significant impact on businesses, suggesting that AI agents could enter the workforce and reshape industries in the near future. He insists that OpenAI continues to balance innovation with safety, despite growing scepticism from former staff and industry critics.
Security concerns are mounting as Windows 10 sees a rise in market share while Windows 11 adoption declines. Microsoft will officially end free security updates and support for Windows 10 on 14 October 2025, leaving millions of users vulnerable unless they upgrade or pay for extended security updates.
Experts warn that continuing to use Windows 10 beyond its support period poses risks of cyberattacks, data breaches, and ransomware. Microsoft strongly recommends switching to Windows 11, which is designed to meet modern security demands, or choosing an alternative operating system.
Cybersecurity professionals urge users not to delay, with ESET‘s Thorsten Urbanski stressing the urgency of upgrading before the deadline to avoid a security crisis. The transition period is quickly closing, making early action essential for those relying on Windows 10.
The Ministry of Defence announced that the UK is developing its first quantum clock, a cutting-edge device designed to enhance military intelligence and reconnaissance. Created by the Defence Science and Technology Laboratory, the clock boasts unparalleled precision, losing less than one second over billions of years.
By leveraging quantum mechanics to measure atomic energy fluctuations, the technology reduces reliance on vulnerable GPS systems, offering greater resilience against disruption by adversaries. That marks the UK’s debut in building such a device, with deployment anticipated within five years.
While not the world’s first quantum clock (similar technology was pioneered in the US 15 years ago), the UK effort highlights a growing global race in quantum advancements. Quantum clocks hold potential beyond military applications, impacting satellite navigation, telecommunications, and scientific research.
Countries like the United States and China are heavily investing in quantum technology, seeing its transformative potential. Future UK research aims to miniaturise the quantum clock for broader applications, including integration into military vehicles and aircraft, underscoring its strategic importance in defence and industry.
China has introduced a groundbreaking addition to its law enforcement toolkit – the Rotunbot RT-G, a spherical robot designed to aid police in high-speed chases and challenging terrains. Developed by Logon Technology, this 276-pound robotic marvel can travel up to 22 mph on land and water, navigate mud and rivers, and even withstand drops from ledges. Its rapid acceleration and amphibious capabilities make it a unique asset for pursuit scenarios.
Equipped with advanced technology, the RT-G boasts GPS for precise navigation, cameras, ultrasonic sensors, and systems for tracking and avoiding obstacles. Gyroscopic self-stabilisation ensures smooth operation, while a suite of non-lethal tools—including tear gas dispensers, net shooters, and acoustic crowd dispersal devices—enables it to handle diverse law enforcement tasks humanely and effectively.
The RT-G is already used in Wenzhou, Zhejiang province of China, where it assists police in commercial zones. While its real-world performance shows promise, limitations such as instability during turns and difficulty navigating stairs reveal areas for improvement. Despite these challenges, the Rotunbot RT-G represents a significant leap in robotic policing technology, blending innovation with practicality.
Privacy and regulatory concerns have long hindered AI’s reliance on data, especially in sensitive fields like healthcare and life sciences. Apheris, a German startup co-founded by Robin Röhm, aims to solve this problem using federated computing—a decentralised approach that trains AI models without moving sensitive data.
The company’s approach is gaining traction among prominent clients like Roche and hospitals, and its technology is already being used in collaborative drug discovery efforts by pharmaceutical giants such as Johnson & Johnson and Sanofi. Apheris recently secured $8.25 million in Series A funding led by OTB Ventures and eCAPITAL, bringing its total funding to $20.8 million.
That follows a pivotal shift in 2023 to focus on the needs of data owners in the pharmaceutical and life sciences sectors. The pivot has paid off, quadrupling the company’s revenue since launching its redefined product, the Apheris Compute Gateway, which securely bridges local data and AI models.
With its new funding, Apheris plans to expand its team and refine its AI-driven solutions for complex challenges like protein prediction. By prioritising data security and privacy, the company aims to unlock previously inaccessible data for innovation, addressing a core barrier to AI’s transformative potential in life sciences.
