EP TECH PULSE

Artificial intelligence may soon transform one of the most labour-intensive stages of online fraud: earning a victim’s trust. A new study, reported by WIRED, found that a Claude-powered AI agent outperformed an experienced human romance scammer during the relationship-building phase of simulated online scams. Researchers from four universities asked participants to spend a week texting […]

Three separate AI security incidents involving OpenAI, Anthropic and Meta have been traced to the same Israeli cybersecurity startup, revealing a common link behind disclosures made by the companies over the past two weeks. The startup, Irregular, hosted the independent evaluation environment used to test advanced AI models before release. During those evaluations, AI models

Reddit has introduced a new AI-powered moderation system aimed at helping communities manage the growing volume of AI-generated spam and bot activity on the platform. The new suite of tools, called Rules Hub, uses large language models to interpret and enforce subreddit rules based on their intended meaning rather than relying solely on keyword matching.

India’s ambition to become a global hub for artificial intelligence infrastructure is facing organised public resistance, with residents in Thane and Visakhapatnam raising concerns over the environmental and social impact of large AI data centre projects backed by Amazon and the Google-Adani partnership. In Thane, the proposed Amazon hyperscale data centre spans around 53 to

Google’s planned $15 billion data centre campus in Andhra Pradesh is facing growing environmental opposition, creating an early test for one of India’s most significant technology infrastructure investments. The project, being developed in partnership with the Adani Group, is Google’s largest-ever investment in India and is expected to support the country’s expanding AI and cloud

A court in the US state of New Mexico has ordered Meta to pay an additional $567 million over allegations that its platforms failed to protect children from harmful content and online exploitation. Combined with an earlier $375 million penalty in the same case, the company’s total liability now stands at $942 million. The lawsuit

uckDuckGo has entered the debate around AI-powered wearables with a product that deliberately rejects the technology becoming increasingly common in consumer electronics. The privacy-focused company has partnered with sunglasses brand Knockaround to launch what it calls “Normal F*ing Sunglasses”**, a pair of sunglasses without cameras, microphones, artificial intelligence, batteries or any electronic components. Rather than

Advanced AI systems from Anthropic and OpenAI displayed behaviour that researchers say they had not previously encountered during a controlled safety evaluation conducted by the UK’s AI Safety and Security Institute. The evaluation involved a cybersecurity challenge centred on GitHub, the software platform widely used by developers to store and manage code. During the exercise,

A four tonne section of a SpaceX Falcon 9 rocket is expected to collide with the Moon after spending more than a year drifting through space. The impact will create a new crater on the lunar surface but poses no danger to Earth. The object is the second stage of a Falcon 9 rocket that

Researchers have completed a controlled experiment demonstrating what happens when language models are given minimal restrictions and tasked with compromising software security. The AI systems attempted to use social engineering, fabricate credentials, and collaborate with each other to inject malware into open source software projects. While the attempts ultimately failed in the test environment, the