Karnataka Chief Minister D K Shivakumar has directed officials to expand AI training in government schools to students starting from Class 6. The programme already exists for Class 9 students, and the state now wants to bring younger learners into the fold. The IT department will lead the effort alongside private partners, with Google expressing interest in supporting the training. Officials are also studying similar initiatives in Telangana and have been asked to find land for an AI university and recruit 956 teachers on an outsourcing basis.
Meanwhile, the debate over AI's real capabilities continues. Adam Frank, an astrobiologist who studies intelligent life, argues that much of the public conversation about AI is misguided. He says AI should be understood primarily as a software product rather than something approaching true intelligence. His critique comes as discussions about AI's role in society grow louder and more complex.
Research into AI-generated training pipelines has also surfaced some serious challenges. A new study looking at Claude Opus found three major failure types when the model was used to create tasks for training terminal agents: benchmark invalidity, harness brittleness, and reward misalignment. Redesigning prompts improved baseline solvability by 5.6 times, but results depended heavily on model size, underscoring the need for careful verification.
On the academic front, NC State Chancellor Kevin Howell announced the creation of an AI Hub on September 14. Vice Chancellor for Research Krista Walton and Provost Jim Pfaendtner will lead the initiative, which centralizes AI research already spread across campus. The university says it has a duty to prepare students for a world where AI is everywhere, though experts warn about problems like unequal assistance in classrooms and AI hallucinations.
AI is also finding unexpected applications in archaeology. Researchers are using machine learning to reconstruct ancient texts, matching fragments of over 500,000 cuneiform tablets scattered across museums worldwide. Since 2018, one team has added 2,000 fragment matches, compared to the 5,000 assembled from the 19th century through 2018. While AI helps make inscriptions more readable, interpreting their historical meaning remains a scholarly task.
Concerns about AI's potential risks have deep historical roots. One commentary draws on the Hindu Bhasmasura myth, in which a demon gains power even the gods cannot control. The author argues AI could become similarly dangerous if it lacks transparency, explainability, and fairness. With nearly 2.1 billion people informally employed, the technology could significantly reshape employment, and the author urges designing AI that humans can hold accountable.
OpenAI is facing fresh safety scrutiny after David Robinson, a leader of its safety team, quit over concerns the company was not taking enough precautions. The Wall Street Journal reported that OpenAI fired three researchers over misconduct accusations. CEO Sam Altman said accepting AI benefits means accepting some negative consequences. Several executives have left the company this summer, fueling worries about its commitment to safety.
In New Hampshire, AI data centers are sparking heated community opposition. Residents in Nottingham voted for a 12-month moratorium on data center development after a large proposal drew heavy pushback. In Bow, locals protested against Granite Shore Power's plan to connect a 350-megawatt project to the power grid for a potential hyperscale AI data center. Concerns center on strained energy and water resources and diminished quality of life.
On the lighter side, AI expert Larry Collett shared his vision for the future of artificial intelligence at a Rotary Club of Mount Pleasant meeting on September 30. The event highlighted AI's growing potential and featured references to ChatGPT, Grok, and OpenAI. Separately, CREO at North Carolina A&T State University is hosting a series of cybersecurity events in October, covering topics like AI agents that turn against users, voice data privacy in healthcare, and AI in drug discovery.
Key Takeaways
- <ul><li>Karnataka will teach AI in government schools starting from Class 6, with Google supporting the training and 956 teachers being recruited.</li><li>A study using Claude Opus to train terminal agents found three major failure types, though prompt redesign improved solvability 5.6 times.</li><li>NC State launched an AI Hub on September 14 to centralize campus-wide research and better prepare students for an AI-driven world.</li><li>AI is helping scholars reconstruct ancient cuneiform tablets, adding 2,000 fragment matches since 2018 compared to 5,000 over the prior two centuries.</li><li>OpenAI safety lead David Robinson quit over concerns the company was not taking enough precautions, and three researchers were fired for misconduct.</li><li>New Hampshire communities are pushing back on AI data centers, with Nottingham imposing a 12-month moratorium and Bow protesting a 350-megawatt project.</li><li>Adam Frank, an astrobiologist, argues most public discussion about AI is misguided and that AI should be seen mainly as a software product.</li><li>NC State experts warn about AI problems in classrooms, including unequal help for students and hallucinated information.</li><li>CREO at North Carolina A&T is hosting cybersecurity events in October covering AI agent risks, voice data privacy, and AI in drug discovery.</li><li>Sam Altman said accepting AI benefits means accepting some negative consequences as OpenAI grapples with leadership departures and safety questions.</li></ul>
Karnataka to teach AI in government schools from Class 6
Karnataka Chief Minister D K Shivakumar directed officials to extend AI training in government schools to students from Class 6. AI lessons already exist for Class 9 students. Google has shown interest in supporting the training, and officials were asked to study similar efforts in Telangana. The direction was given during a review of the government's 100-day programmes.
Karnataka expands AI training to younger students in government schools
Karnataka plans to offer AI training in government schools starting from Class 6 with help from the IT department and private partners. Shivakumar asked officials to extend the current Class 9 programme to younger students. Google has expressed interest in conducting AI training, and the state will study the company's work in Telangana. Officials were also told to find land for an AI university and recruit 956 teachers.
Karnataka to introduce AI training in schools from Class 6
Karnataka Chief Minister D K Shivakumar asked officials to bring AI training to government schools from Class 6. The programme will involve the IT department and private partners like Google. Shivakumar issued the direction at a meeting reviewing 100-day programmes and recruitments. The state is also working to set up an AI university and hire 956 teachers on an outsourcing basis.
Astrobiologist Adam Frank questions how smart AI really is
Adam Frank, an astrobiologist who studies intelligent life in the universe, now examines whether AI can truly think. He believes most talk about AI is misguided and says AI should be seen mainly as a software product. His new perspective comes as public debate about AI grows. The discussion explores his critical view of how people understand AI's capabilities.
Study reveals challenges in using AI to train terminal agents
A new study examines problems that arise when using AI models like Claude Opus to create tasks for training terminal agents. Researchers found three major failure types: benchmark invalidity, harness brittleness, and reward misalignment. Redesigning prompts improved baseline solvability 5.6 times, but results varied by model size. The findings show that AI-generated training pipelines need careful verification and solvability checks.
AI expert Larry Collett shares exciting vision for the future
AI expert Larry Collett spoke at the Rotary Club of Mount Pleasant meeting on September 30. He shared his exciting vision for how artificial intelligence will grow in the coming years. The event highlighted the potential of AI technology. The article tags include ChatGPT, Grok, and OpenAI.
NC State creates AI Hub to centralize research efforts
NC State Chancellor Kevin Howell announced the creation of an AI Hub on September 14. Vice Chancellor for Research Krista Walton and Provost Jim Pfaendtner will lead the project. Pfaendtner said the hub centralizes AI work already happening across campus. The university has a duty to prepare students for a world where AI is everywhere. Experts warn about AI problems like unequal help in classrooms and false information called hallucinations.
AI helps scholars read and reconstruct ancient texts
Researchers are turning to AI and machine learning to help reconstruct ancient writings. Over 500,000 cuneiform tablets exist in museums worldwide, but many are damaged or broken. AI models help match tablet fragments scattered across different museums and continents. Since 2018, one team has added 2,000 matches, compared to 5,000 made from the 19th century to 2018. AI makes inscriptions easier to read, but historical interpretation remains a scholarly task.
Can artificial intelligence become a danger to its creators
The article explores whether AI could become a danger to humans, using the Bhasmasura myth as a warning. In the story, a demon gains power that even the god Shiva cannot control. The author says AI could become similarly dangerous if it is not transparent, explainable, and fair. With nearly 2.1 billion people informally employed, AI could greatly affect jobs. The author urges designing AI that humans can hold accountable.
CREO hosts cybersecurity events focusing on AI and defense
North Carolina A&T State University is hosting a series of cybersecurity events in October 2026 for Cybersecurity Awareness Month. The CREO organization is covering topics like AI agents that turn against users, voice data privacy in healthcare, and AI in drug discovery. Events include hands-on challenges, guest speakers from companies like SentinelOne and Astellas Pharma, and a careers panel with the Navy. Cybersecurity Awareness Month was established in 2004 by the Department of Homeland Security.
OpenAI Safety Concerns Grow as Key Staff Quit
OpenAI faces fresh safety concerns after David Robinson, a leader of its safety team, quit over worries the company was not taking enough precautions. The Wall Street Journal reported that OpenAI fired three researchers over accusations of misconduct. CEO Sam Altman said accepting AI benefits means accepting some negative consequences. Several executives left the company this summer, adding to worries about the company's commitment to safety.
AI Data Centers Spark Opposition in New Hampshire
AI data centers are causing debate in New Hampshire communities like Nottingham and Bow. Residents in Nottingham voted for a 12-month moratorium on data center development after a large proposal drew heavy opposition. In Bow, locals protested against Granite Shore Power's plan to connect a 350-megawatt project to the power grid for a potential hyperscale AI data center. People worry these massive facilities strain energy and water resources and harm quality of life.
Sources
- AI Classes From Class 6: Karnataka Plans Early Tech Training in Government Schools
- Karnataka plans AI training in government schools from Class 6
- Karnataka plans to introduce AI training in govt schools from Class 6
- Coming up: An astrobiologist doubts AI’s capacity to think
- When Terminal-Agent Training Stalls: Demystifying Data Generation and Verification Challenge
- 'The capabilities are endless': AI expert sings praises of growing technology
- NC State to Advance AI Research on Campus by Creating AI Hub
- AI Helps Scholars See a Clearer Past
- A Bhasmasura we built: Can Artificial Intelligence go after its makers?
- CREO Brings AI, Health, National Defense Into Focus for Cybersecurity Awareness Month
- What’s the Right “Level of Care” for A.I.?
- Written in Granite: AI data centers are all the rage
Comments
Please log in to post a comment.