large language models

The Comptroller and Auditor General (CAG) is developing a sovereign AI platform, including a large language model, to enhance public sector auditing. This initiative aims to leverage AI and ML tools for faster and more efficient identification of procurement concentration and cartel-risk indicators by analyzing vast datasets and detecting systemic patterns.

A groundbreaking UC San Diego study reveals advanced AI can now pass the Turing test, convincingly mimicking human conversation, tone, and even errors. Researchers found participants struggled to distinguish between humans and sophisticated AI models like GPT-4.5, raising concerns about online trust and the potential for AI deception.

Nvidia's Jonathan Ross suggests job applicants can boost their chances by using AI models similar to those recruiters employ for resume screening. Research indicates AI hiring systems may favor resumes generated by their own underlying large language models, potentially increasing shortlisted candidates' probability by up to 60%.

IBM is investing $5 billion in Project Lightwell, a cybersecurity initiative to shield open-source software from advanced AI threats. Driven by concerns over AI models like Anthropic's Mythos, the project mobilizes 20,000 engineers from IBM and Red Hat. Major banks like Goldman Sachs and JPMorgan are already on board to bolster their defenses against emerging vulnerabilities.

Chinese AI startup DeepSeek has slashed the price of its V4-Pro model by 75%, making advanced AI integration significantly cheaper for developers. This drastic price reduction is attributed to the increased availability of Huawei's Ascend 950 chips, allowing DeepSeek to optimize its software on local silicon and bypass costly, restricted Nvidia hardware.

China is enhancing its nationwide surveillance with AI and LLM-powered systems, shifting towards predictive policing. Local governments are deploying new technologies that interpret video feeds and retrieve footage via text prompts, automating monitoring and identifying unusual behavior in real-time. This upgrade aims to improve efficiency and proactive risk assessment.

American scientist Gary Marcus has hit back at Oracle founder Larry Ellison's claim that AI models like ChatGPT, Gemini, Grok, and Llama are commoditised because they train on the same public data. Marcus says he warned of this exact "no moat" problem two years ago, predicting price wars and weak differentiation. He argues Silicon Valley's refusal to listen will cost the industry billions.

Shark Tank investor Mark Cuban warns OpenAI's massive fundraising may lead to 'deep trouble' if economics don't deliver, questioning AI infrastructure spending returns. He also expressed uncertainty about the LLM market's future and criticized AI firms for hyping risks. Separately, Cuban revealed selling most Bitcoin, finding it an unreliable hedge against global turmoil.

Airbnb CEO Brian Chesky defended the company's use of Chinese AI models, stating US lawmakers misunderstand open-source systems. He clarified that Airbnb does not share user data with Chinese firms, primarily utilizing open-source models, including those from the US. This defense comes amid congressional concerns over data security and potential national security risks.

AI firm Anthropic is reportedly in early discussions with Microsoft to rent server space powered by the tech giant's custom-designed AI processors. This move aims to diversify Anthropic's hardware backend and reduce reliance on Nvidia, following a significant infrastructure deal with Google. The potential partnership underscores a broader industry trend of reducing dependence on Nvidia's dominant AI chips.

India's student startups are pivoting from consumer internet to deep-tech, with AI, defence, and climate leading the charge. This shift signifies a rise in technical sophistication, as evidenced by a surge in PhD founders and a widening geographical spread of innovation beyond major metros. Edtech and services sectors have seen a notable decline.

A Times Techies discussion in partnership with Accenture brought together senior leaders from ANSR, Inspire Brands and Deutsche Boerse to examine what it will take for India’s global capability centres to lead, not merely support, the AI-driven enterprise of tomorrow.

Palantir CTO Shyam Sankar challenges the narrative of AI-driven job losses, arguing that frontline workers, not tech executives, offer the true perspective. He highlights how automation in manufacturing and healthcare is enhancing efficiency and enabling job creation, rather than widespread displacement. Sankar emphasizes listening to those directly using AI in real-world applications.

Tech giants like Uber are rethinking their massive AI investments. Concerns are rising over high costs and the difficulty in proving AI's real-world benefits. Language learning app Duolingo is also stepping back from AI usage metrics. Research suggests that more AI usage does not automatically mean better results. Companies now focus on efficiency and return on investment.

During the convocation at Nalanda University, P.K. Mishra, Principal Secretary to the Prime Minister, underscored the pivotal role of Asia in global affairs alongside India’s robust economic position. He championed a model of sustainable development that integrates knowledge, values, and harmonious living.

Google I/O 2026 kicks off at Shoreline Amphitheatre, promising major AI advancements and operating system updates. Expect next-generation Gemini models with agentic capabilities, a potential Android-ChromeOS merger for new "Googlebooks" laptops, and a closer look at augmented reality smart glasses. Creative AI tools like Veo and Lyria will also be showcased.

Europe faces a critical two-year deadline to build its own AI infrastructure. Mistral AI's Arthur Mensch warns of US technology dependence if Europe fails to act. Control over chips, energy, and computing power is key to AI competition. Failure could leave Europe reliant on American digital services. Mistral is working to expand Europe's AI capacity.

Anthropic and OpenAI have fired another salvo, launching their own services companies. While some believe this will force software firms to become more differentiated, others price these stocks at zero earnings growth

Take-Two Interactive CEO Strauss Zelnick believes AI can generate GTA lookalikes but not original hits, stating that clones don't sell. He emphasized that while AI excels at asset creation and improving efficiency, it lacks the originality and creative thinking essential for culturally influential games. Zelnick also dismissed concerns about AI democratizing game development, noting that technology has always been accessible.

Google's SynthID technology embeds invisible watermarks into AI-generated content, making it detectable even after modifications. Announced at Google I/O 2026, SynthID is now integrated into Search and Chrome, simplifying verification for users. This expansion aims to combat the growing challenge of distinguishing real from synthesized media.

Google DeepMind CEO Demis Hassabis has called AI-driven layoffs "dumb," telling Wired ahead of Google I/O that companies cutting engineers to chase AI productivity gains are showing a "lack of imagination." His comments come as Anthropic's Dario Amodei warns of mass white-collar job losses, and as Meta, Amazon, Block and others blame AI for thousands of cuts.

Mark Cuban warns that AI's inconsistency makes human expertise crucial, debunking 'doomer' fears of AI takeover. He highlights that AI's probabilistic nature means it doesn't grasp consequences, making human judgment indispensable. Meanwhile, Google and Uber CEOs reveal AI generates significant code, but human engineers always provide final approval, underscoring the continued need for human oversight.

The anticipated National School Mental Health Policy in India marks a pivotal moment in aligning student wellness with educational success. By fostering nurturing atmospheres, this initiative seeks to empower 40 million learners, promoting empathy and resilience. To realize this vision, effective teacher training and adaptable strategies will be key, heralding a hopeful era for the youth of the nation.

Cybercriminals have reportedly leveraged artificial intelligence to craft a potent hacking tool capable of circumventing security measures in a widely used system administration software. Google's security team uncovered this novel threat, which exploited an unknown vulnerability to bypass multi-factor authentication. The developer was alerted and patched the flaw before it could be weaponized, marking a significant development in AI-driven cyberattacks.