Two fully produced AI-generated sci-fi series, Space Nation and Kira, premiered at the Mipcom TV market in Cannes. Ex Machina Studios produced both shows, with K5 Intl. handling worldwide sales. Space Nation consists of eight 45-minute episodes, while Kira has six 35-minute episodes. Globoplay joined as co-producer and holds Brazilian rights for Space Nation. The pilots screened for buyers on October 12 and 13, marking the first time fully AI-generated series episodes were shown at a major international market.
Amazon plans a $21 billion investment in AI infrastructure in India by 2030, but the expansion is drawing pushback from local residents. Near Mumbai, trees were cleared to make way for a data center, sparking tension between national AI ambitions and local impact. Prime Minister Modi has pushed for India to lead in the AI age, yet communities near data center sites are raising concerns about the effects of the building boom.
Keith Sullivan, CTO of AFL, discussed the challenges of building AI training clusters in a DCD Talks episode. He stressed the importance of collaboration, data quality, and strong infrastructure for AI workloads. Sullivan also shared AFL's experience using NVIDIA DGX A100 systems and cloud-based services for AI projects.
Meta donated AI-powered glasses to blind and visually impaired veterans in Des Moines, Iowa. At least ten veterans received the voice-controlled devices along with hands-on training. The glasses can read texts, identify objects, and help users navigate their surroundings. The program is part of Meta's plan to provide AI glasses to every legally blind veteran in the United States.
New research from LMU Munich introduces byteification, a method that converts existing AI language models into byte-level models using less than one percent of the typical training required. The resulting models can read individual letters instead of only word chunks and outperform previously published byte-level models of similar size. The models, code, and data are publicly available.
A PwC survey of 49,364 workers across 48 countries found that AI use rose 10 percentage points in one year, reaching 64 percent of workers. However, access to training fell eight points to 51 percent in 2026. The report divides workers into front-runners and engine room groups, with the engine room having less access to learning resources and lower confidence in job security.
Driven Tech joined the Open Secure AI Alliance, a Linux Foundation initiative focused on AI security. The company submitted a Request for Comments to help govern and secure AI-enabled systems, with a focus on securing agentic systems, AI workflows, and autonomous technologies. Executives Sanjeev Pant and Jeff Ridge said the alliance helps turn AI security into practical, measurable outcomes.
Researchers tested seven open-weight language models on legal reasoning tasks and found that naming a legal authority does not reliably indicate a verdict depends on it. When the cited authority was changed, the verdict often stayed the same. The findings raise concerns about relying on AI-generated legal explanations for compliance or audit purposes.
The Trump administration launched America.gov, an AI-powered search engine for federal information and services. Lafayette College professors raised concerns about reliability, privacy, environmental costs, and whether the site addresses deeper problems with government access. They debated whether the chatbot reduces errors compared with existing government websites and stressed the need for human recourse for users.
MIT Technology Review hosted a roundtable on October 16, 2026 about AI-designed viruses. Samuel King, a bioengineering PhD candidate at Stanford University and the Arc Institute, joined the AI reporter to discuss whether AI can design new life forms. The conversation also explored how today's AI still lacks the creativity needed for truly innovative, open-ended research.
Key Takeaways
- Space Nation and Kira, two fully AI-generated sci-fi series, premiered pilot episodes at Mipcom in Cannes, sold by K5 Intl. and produced by Ex Machina Studios.
- Amazon plans a $21 billion investment in AI infrastructure in India by 2030, facing local pushback over data center expansion near Mumbai.
- AFL CTO Keith Sullivan discussed building AI training clusters using NVIDIA DGX A100 systems in a DCD Talks episode.
- Meta donated AI-powered glasses to blind veterans in Des Moines as part of a plan to reach every legally blind veteran in the United States.
- LMU Munich researchers developed byteification, a method that turns AI language models into byte-level models using less than one percent of normal training.
- A PwC survey of 49,364 workers found AI use at 64 percent but training access dropped to 51 percent, widening the gap between front-runners and engine room workers.
- Driven Tech joined the Open Secure AI Alliance and submitted an RFC focused on securing agentic AI systems and autonomous workflows.
- Study finds AI language models citing legal authorities often produce unchanged verdicts even when those authorities are altered, questioning reliability for legal compliance.
- America.gov, an AI-powered federal search engine, launched under the Trump administration amid concerns from Lafayette College professors about reliability and privacy.
- MIT Technology Review roundtable explored AI-designed viruses, with Stanford's Samuel King noting AI still lacks creativity for open-ended biological research.
AI-made series Space Nation and Kira debut at Mipcom
Two fully produced AI-generated sci-fi series, Space Nation and Kira, premiered pilot episodes at the Mipcom TV market in Cannes. Both shows are produced by Ex Machina Studios and sold by K5 Intl. Space Nation has eight 45-minute episodes and Kira has six 35-minute episodes. Globoplay joined as co-producer and holds Brazilian rights for Space Nation. The pilots screened for buyers on October 12 and 13.
AI series Space Nation and Kira screen at Mipcom market
Space Nation and Kira, two science-fiction series made with generative AI, had their world premieres at Mipcom in Cannes. Ex Machina Studios produced both shows, with K5 Intl. handling worldwide sales. The companies said this is the first time fully produced AI-generated series episodes were screened for buyers at a major international market. The pilots were shown on October 12 and 13 at the Cinéma Les Arcades.
India's AI data center expansion faces local pushback
Indian engineer Tushar Shetty watched lush greenery near his Mumbai apartment replaced by red earth as trees were cleared for an Amazon data center. Amazon plans a $21 billion investment in AI infrastructure in India by 2030. Prime Minister Modi wants India to lead in the AI age, but residents are pushing back against the data center building boom. The tension between AI growth and local impact is playing out around the world.
New method lets AI language models read every letter
Researchers from LMU Munich and other institutions developed a method called byteification that converts existing AI language models into byte-level models. The process requires less than one percent of the training typically needed. This lets models see individual letters instead of only word chunks. The resulting models outperform all previously published byte-level AI language models of similar size. The models, code, and data are publicly available.
Driven Tech joins Open Secure AI Alliance
Driven Tech joined the Open Secure AI Alliance, a Linux Foundation initiative focused on AI security. The company submitted a Request for Comments to help govern and secure AI-enabled systems. Driven's RFC focuses on securing agentic systems, AI workflows, and autonomous technologies. Executives Sanjeev Pant and Jeff Ridge said the alliance helps turn AI security into practical, measurable outcomes. Driven also plans to contribute expertise in AI governance and risk management.
DCD Talks explores building AI training clusters with AFL CTO
Keith Sullivan, CTO of AFL, discusses the challenges of building AI training clusters in a DCD Talks episode. He highlights the importance of collaboration, data quality and strong infrastructure for AI workloads. He also shares AFL's experience using NVIDIA DGX A100 systems and cloud-based services for AI.
AI adoption rises but worker training lags behind, report finds
A PwC survey of 49,364 workers across 48 countries found AI use rose 10 percentage points in one year, with 64 percent of workers now using AI. However, access to training fell eight points to 51 percent in 2026. The report splits workers into front-runners and engine room groups, with the engine room having less access to learning resources and lower job security confidence.
Meta donates AI glasses to blind veterans in Des Moines
Blind and visually impaired veterans in Des Moines, Iowa received AI-powered glasses from Meta. At least ten veterans got the voice-controlled glasses along with hands-on training. The glasses can read texts, identify objects and help users navigate their surroundings. The program is part of Meta's plan to provide AI glasses to every legally blind veteran in the United States.
Study tests whether legal AI reasoning truly follows cited authorities
Researchers tested seven open-weight language models from 8B to 70B parameters on legal reasoning tasks. When models named a legal authority to justify a verdict, changing that authority often did not change the verdict. The study found that naming a legal authority is a poor indicator that a verdict actually depends on it. The findings raise concerns about using AI-generated legal explanations as compliance or audit tools.
Lafayette professors weigh in on government AI chatbot America.gov
The Trump administration launched America.gov, an AI-powered search engine for federal information and services. Lafayette College professors raised concerns about reliability, privacy, environmental costs and whether the site addresses deeper problems with government access. Professors debated whether the chatbot reduces errors compared with existing government websites and stressed the need for human recourse for users.
Inside the Race to Design Viruses Using AI
MIT Technology Review hosted a roundtable on October 16, 2026 about AI-designed viruses. Samuel King, a bioengineering PhD candidate at Stanford University and the Arc Institute, joined the AI reporter to discuss the topic. The conversation explored whether AI can design new life forms. It also touched on how today's AI still lacks the creativity for truly innovative open-ended research.
Sources
- AI-Produced Series ‘Space Nation’ and ‘Kira’ to Premiere at Mipcom
- AI-Produced Series ‘Space Nation’ and ‘Kira’ to Premiere Pilot Episodes at Mipcom (EXCLUSIVE)
- Modi wants India to power the AI age. Residents are pushing back
- Artificial Intelligence: Language Models That See Every Letter
- Driven Tech Joins Open Secure AI Alliance and Contributes AI Security RFC
- DCD Talks: Building AI training clusters with Keith Sullivan, AFL
- AI use is up, but worker training is falling behind: report
- We The People: AI glasses for veterans
- Cited but Not Consulted: A Counterfactual Audit of Legal Chain-of-Thought Faithfulness
- Professors on the government’s new AI-chatbot
- Roundtables: A Conversation With the Creator of AI-Designed Viruses
Comments
Please log in to post a comment.