Two cybersecurity firms have tested OpenAI’s newly released GPT-5, revealing that the AI model remains vulnerable to manipulation despite its advanced safety features. According to separate analyses by NeuralTrust and SPLX, specialised techniques can bypass GPT-5’s guardrails, raising concerns about its readiness for high-stakes enterprise use.
Google has introduced a 'Guided Learning' feature for its Gemini AI platform, arriving just over a week after OpenAI's July 29 launch of a similar 'Study Mode' for ChatGPT. Both tools provide structured learning assistance rather than direct answers, marking a shift in how AI platforms approach educational support.
If you’ve ever gone down a rabbit hole with ChatGPT, you might soon get a gentle nudge to take a break. OpenAI is rolling out a new feature that reminds users when their chat sessions start running too long, as a part of a broader effort to help people stay in control of their time.
A new study has found that ChatGPT, Google’s AI Overviews, and Perplexity AI cite wildly different sources of information, raising important questions for brands and content creators trying to stay visible in the age of AI.
OpenAI has launched a new 'study mode' for ChatGPT, intended to guide students through learning rather than directly providing answers. The feature, now available to all users, including free accounts, will soon be extended to ChatGPT Edu, a specialised version of the AI tool for educational institutions.
OpenAI CEO Sam Altman has recently raised concerns about privacy protections for sensitive ChatGPT conversations, comparing today's AI privacy standards to "having therapy sessions in public". Speaking on an interview on the 'This Past Weekend' podcast, Altman acknowledged growing user hesitation about sharing personal information with AI tools.
OpenAI has announced a $50 million (£39 million) fund to help nonprofit and community organisations implement AI solutions. The fund will support organisations working in education, healthcare, economic opportunity, and community organising, says the company.
OpenAI has expanded its computing power sources by partnering with Google Cloud, according to an update on its website. The deal, finalised in May after months of negotiations, marks a strategic shift as the ChatGPT maker diversifies beyond its primary backer, Microsoft.
Google has pulled off a strategic talent acquisition by hiring key executives and engineers from AI coding startup Windsurf, following failed acquisition talks between Windsurf and Google rival OpenAI. Google will pay $2.4 billion in licensing fees to access some of Windsurf's technology while allowing the startup to remain independent, according to a report by Reuters.
Two cybersecurity firms have tested OpenAI’s newly released GPT-5, revealing that the AI model remains vulnerable to manipulation despite its advanced safety features. According to separate analyses by NeuralTrust and SPLX, specialised techniques can bypass GPT-5’s guardrails, raising concerns about its readiness for high-stakes enterprise use.
Google has introduced a 'Guided Learning' feature for its Gemini AI platform, arriving just over a week after OpenAI's July 29 launch of a similar 'Study Mode' for ChatGPT. Both tools provide structured learning assistance rather than direct answers, marking a shift in how AI platforms approach educational support.
If you’ve ever gone down a rabbit hole with ChatGPT, you might soon get a gentle nudge to take a break. OpenAI is rolling out a new feature that reminds users when their chat sessions start running too long, as a part of a broader effort to help people stay in control of their time.
A new study has found that ChatGPT, Google’s AI Overviews, and Perplexity AI cite wildly different sources of information, raising important questions for brands and content creators trying to stay visible in the age of AI.
OpenAI has launched a new 'study mode' for ChatGPT, intended to guide students through learning rather than directly providing answers. The feature, now available to all users, including free accounts, will soon be extended to ChatGPT Edu, a specialised version of the AI tool for educational institutions.
OpenAI CEO Sam Altman has recently raised concerns about privacy protections for sensitive ChatGPT conversations, comparing today's AI privacy standards to "having therapy sessions in public". Speaking on an interview on the 'This Past Weekend' podcast, Altman acknowledged growing user hesitation about sharing personal information with AI tools.
OpenAI has announced a $50 million (£39 million) fund to help nonprofit and community organisations implement AI solutions. The fund will support organisations working in education, healthcare, economic opportunity, and community organising, says the company.
OpenAI has expanded its computing power sources by partnering with Google Cloud, according to an update on its website. The deal, finalised in May after months of negotiations, marks a strategic shift as the ChatGPT maker diversifies beyond its primary backer, Microsoft.
Google has pulled off a strategic talent acquisition by hiring key executives and engineers from AI coding startup Windsurf, following failed acquisition talks between Windsurf and Google rival OpenAI. Google will pay $2.4 billion in licensing fees to access some of Windsurf's technology while allowing the startup to remain independent, according to a report by Reuters.
Meta Platforms, the parent company of Facebook, WhatsApp, and Instagram, is reportedly hiring four additional artificial intelligence researchers from OpenAI, according to a Saturday report by The Information. The latest recruits—Shengjia Zhao, Jiahui Yu, Shuchao Bi, and Hongyu Ren—will join Meta's AI division as the company expands its artificial intelligence capabilities.
