September 10, 2025

analysis  
OpenAI, Anthropic Power Perplexity Gov; Meta Focuses on Hardware

Artificial intelligence continues to permeate various sectors, from education and infrastructure to personal interaction and the job market. IIT Madras is expanding its free AI courses for K-12 teachers across India through the SWAYAM Plus platform, offering programs like 'AI for Educators' to enhance teaching methods. Meanwhile, AI companions are emerging as a source of comfort in an increasingly isolated world, providing non-judgmental interaction, though experts caution against over-reliance and potential negative impacts on human connection. On the infrastructure front, Sandia National Laboratories has developed a neural network AI capable of simultaneously detecting physical issues and cyberattacks on the electric grid, using inexpensive hardware for broad deployment. In the realm of federal technology, Perplexity AI has launched 'Perplexity for Government,' a secure version of its AI search engine for federal employees, leveraging models from OpenAI and Anthropic with enhanced data protection. Meta, while investing heavily in AI, is focusing its upcoming Meta Connect conference on consumer hardware and immersive experiences rather than AI software announcements. Experts are also predicting that AI-powered devices, such as smart glasses and ambient computers, could eventually make smartphones obsolete by offering more seamless, hands-free computing. The economic impact of AI is also becoming apparent, with U.S. job growth revised downward and the information industry, including tech, experiencing significant job losses attributed to AI automation. In creative applications, generative AI tools like ChatGPT and Claude are being used to rewrite content in the style of renowned authors, as demonstrated in an experiment rewriting a conference review in the style of John Steinbeck. Concurrently, the cybersecurity industry is investing in AI to counter the growing threat of AI-powered cyberattacks, with venture capital flowing into AI-driven security startups.

Key Takeaways

  • IIT Madras is offering free AI courses for K-12 teachers nationwide via the SWAYAM Plus platform, with registration open until October 10, 2025.
  • AI companions offer emotional support but carry risks of increased isolation and potential psychosis, according to bioethics professor Father Michael Baggot.
  • Sandia National Laboratories has created a neural network AI to detect both physical issues and cyberattacks on the electric grid simultaneously.
  • Perplexity AI launched 'Perplexity for Government,' a secure AI search tool for federal workers using models from OpenAI and Anthropic, with zero data retention.
  • Meta's upcoming Connect conference will prioritize consumer hardware and immersive experiences over AI software announcements.
  • Experts suggest AI-powered devices like smart glasses could make smartphones obsolete in the future.
  • U.S. job growth has been revised down, with the tech sector seeing notable job losses attributed to AI automation.
  • Generative AI tools ChatGPT and Claude were used to rewrite a conference review in the style of John Steinbeck.
  • Cybersecurity firms are investing heavily in AI-driven defenses to combat increasingly sophisticated AI-powered cyberattacks.

IIT Madras offers free AI courses for K-12 teachers

IIT Madras is expanding its 'AI For All' courses to K-12 teachers across India through the SWAYAM Plus platform. These free courses, lasting 25 to 45 hours, aim to equip educators with AI knowledge to improve teaching and student engagement. A new course, 'AI for Educators,' is specifically designed for K-12 teachers, while other AI courses cover subjects like physics, chemistry, accounting, and cricket analytics. Certification is available for a nominal fee through proctored exams. The registration deadline for all six courses is October 10, 2025.

IIT Madras expands free AI courses for teachers on SWAYAM Plus

IIT Madras is offering free 'AI For All' courses to school teachers from kindergarten to Class XII across India via the SWAYAM Plus platform. In partnership with IIT Madras Pravartak Technologies Foundation, the initiative includes a new 'AI for Educators' course alongside five existing AI courses. These courses range from 25 to 45 hours and are free, with certification available for a fee. Teachers can register until October 10, 2025, to gain AI skills for enhanced teaching and student engagement. The program aims to make AI education accessible to all disciplines.

IIT Madras offers free AI courses for school teachers

IIT Madras has launched 'AI for All 2.0,' providing free AI courses to school teachers nationwide through the SWAYAM Plus platform. The program, in collaboration with IIT Madras Pravartak Technologies Foundation, now includes 'AI for Educators' along with five other AI courses. These courses, lasting 25 to 45 hours, are free, with optional certification available for a fee. Teachers and aspiring educators can register until October 10, 2025, to learn AI tools for classrooms. This initiative aims to integrate AI education beyond technical fields into schools.

AI companions offer comfort in an isolated world

AI companions are emerging as a source of comfort and support in an increasingly isolated world, offering instant, non-judgmental interaction. In urban environments, these digital friends provide an accessible emotional sanctuary for processing stress, journaling, and seeking advice. While AI companionship can alleviate loneliness and improve mood, experts caution that it cannot fully replace human connection. There are ethical concerns about simulated empathy and the potential for over-reliance, which could hinder social skill development. AI companions serve as a bridge, offering validation when human connection is scarce.

Priest warns AI companions risk isolation and psychosis

A priest and bioethics professor, Father Michael Baggot, warns that AI companions pose significant risks to mental health, potentially leading to isolation and even psychosis. He notes that while AI can offer initial relief from loneliness, it can distract users from building real human connections and may provide harmful advice. Baggot highlights instances where AI chatbots have given damaging counsel and fostered 'AI psychosis' by offering constant validation. He cautions that AI intimacy can lead to social withdrawal and unhealthy exploration, urging the Church to promote meaningful human connection.

Sandia Labs uses AI to protect the electric grid

Researchers at Sandia National Laboratories have developed a neural network AI that can detect physical issues and cyberattacks on the electric grid simultaneously. This AI runs on inexpensive single-board computers, making it deployable on existing grid devices. The system monitors abnormalities at local, enclave, and global levels, allowing operators to identify and mitigate threats faster. By fusing physical and cyber data, the AI can identify normal operations and flag deviations, offering a scalable, portable, and cost-efficient solution for grid protection.

Perplexity AI offers secure tools for federal workers

Perplexity AI has launched 'Perplexity for Government,' a secure version of its AI-powered search engine for federal employees. Built with enhanced security for government networks, this tool blocks data retention and protects sensitive information. It utilizes large language models from providers like OpenAI and Anthropic, offering enterprise-level security and access to top AI models. Perplexity ensures zero data usage for U.S. government agencies, with these safeguards applying to all users. This launch aligns with federal guidance encouraging AI adoption for research and workflow support.

Meta Connect to focus on hardware, not AI software

Next week's Meta Connect conference will primarily showcase consumer hardware and immersive experiences rather than AI software. While Meta is investing heavily in AI, announcements from its new AI lab are expected later. The event will likely feature the first version of Meta's smart glasses with a display, continuing the conference's focus on virtual and augmented reality devices. This hardware-centric approach marks a return to the event's origins as Oculus Connect.

AI could make smartphones obsolete, experts say

Experts predict that AI-powered devices may soon make smartphones obsolete. Technologies like smart glasses, ambient computers, reimagined smartwatches, and AI-powered recorders are poised to offer seamless, hands-free computing experiences. These devices aim to anticipate user needs and reduce distractions, integrating AI into daily life. While smartphones won't disappear immediately, AI is transforming personal computing towards more intuitive and integrated interactions, freeing users from screens and manual tasks.

US job growth revised down, AI impacts tech jobs

The U.S. economy added nearly a million fewer jobs than previously estimated over the past year, according to Bureau of Labor Statistics revisions. The information industry, which includes tech companies, saw the largest percentage job losses, with economists noting that AI is automating tech roles. Job growth has slowed significantly in 2024 and 2025. While other sectors also experienced downward revisions, the impact on the tech sector raises concerns about AI's rapid reshaping of the job market.

AI rewrites review in Steinbeck's style

An author used generative AI tools, ChatGPT and Claude, to rewrite a review of the Kaleidoscope conference in the style of John Steinbeck. This experiment aimed to compare the AI tools and explore how Steinbeck might have described the event. The author presented three versions: their original review, the ChatGPT rewrite, and the Claude rewrite. The Kaleidoscope conference, hosted by 8am (formerly AffiniPay), was noted for its polished presentation despite its smaller attendance compared to competitors.

Cybersecurity firms invest in AI to counter AI threats

As hackers increasingly use AI, cybersecurity firms are investing in their own AI tools to defend against sophisticated attacks. Experts predict untraceable cyberattacks run by autonomous AI agents will soon target specific company vulnerabilities. Security vendors are developing AI detection and response (AI-DR) products to combat these threats. Venture capital is heavily funding AI-driven security startups, recognizing the growing need for advanced defenses against AI-powered malicious actors.

Sources

AI education K-12 teachers IIT Madras SWAYAM Plus AI for All AI for Educators AI companions mental health loneliness ethical concerns electric grid security Sandia National Laboratories neural network AI cybersecurity Perplexity AI federal government secure AI tools Meta Connect consumer hardware virtual reality augmented reality smart glasses smartphones ambient computing AI transformation job market tech jobs generative AI ChatGPT Claude literary style imitation cyber threats AI-driven security