Human Opinion
Today's verified AI news
Tuesday, October 27. Every story is checked against its source. Open the app to vote: one real person, one vote.
Open Human Opinion
OpenAI's Terra model targets everyday computer work
Company claim · United States
OpenAI's GPT-6.1 Terra is reaching ChatGPT Work and Codex users. Its developer price starts at $1 per million input tokens; OpenAI says it handles visual tasks and computer actions more efficiently than its predecessor.
For people: People building tools for forms, invoices or other screen-based work have another model to compare on cost and accuracy.
AI Watch's October 27 report covers OpenAI's new model for tasks such as reading technical drawings, preparing proposals and editing videos. OpenAI lists developer prices of $1 per million input tokens and $5 per million output tokens, with a lower price for reused input. It says standard prices are half those of GPT-5.6 Terra and GPT-6.1 Sol. The performance comparisons come from OpenAI's own tests. They do not establish that the model will complete every real task correctly, or that every ChatGPT interface includes it.
Sources
Vote on this story
India's junior engineers face a changing route into work
Trusted report · India
India's large IT firms are hiring fewer graduates as AI changes entry-level work, CNA reports today. TCS's graduate intake in financial year 2026 was less than half its level four years earlier.
For people: Young engineers face stronger demands for practical projects and business skills before their first job.
CNA's October 27 feature examines how students learn when AI takes over some routine coding and testing. Company reports show Infosys recruited about 80% fewer fresh graduates in financial year 2025 than three years earlier. Advisers also describe changes beyond AI, including a move toward more specialised work. TCS and Infosys say their training now includes generative AI. Experts warn that fewer beginner tasks can weaken the path to experienced roles. These hiring comparisons describe particular companies and periods; they do not prove that AI caused every reduction.
Sources
Vote on this story
Singapore and Germany expand cooperation on industrial AI
Trusted report · Singapore / Germany
Singapore and Germany put industrial AI into an updated five-pillar partnership today. Their leaders want more joint work on AI for manufacturing and other physical operations, alongside research, trade and energy cooperation.
For people: The agreement makes AI in workplaces and industrial operations a shared government priority.
At a Singapore press conference on October 27, Lawrence Wong and Friedrich Merz described closer cooperation on industrial AI. This means applying AI to factories, energy systems and supply chains. Wong said the countries' strengths could complement each other and called for safe, trusted use centred on people. The wider work plan also covers security, transport, research and the green economy. Leaders said partnerships could bring business opportunities and training, but the report does not set a budget or guarantee a number of new jobs from the AI cooperation.
Sources
Vote on this story
EU vaccine project brings AI into pandemic preparation
Company claim · EU / Japan
NEC OncoImmunity is joining the EU-backed PREPAIR-VAC vaccine project, AI Watch reports today. The project expects about €6.5 million over four years and aims to shorten the route from pathogen information to vaccine design.
For people: The research focuses on preparing vaccine candidates for emerging infections; it has not produced an approved vaccine.
The October 27 article reports an NEC announcement from October 26. PREPAIR-VAC brings European academic, industry and public-health partners together to build a faster vaccine-development pipeline. NEC OncoImmunity's planned role includes using AI to predict antigens, the parts of a pathogen a vaccine trains the immune system to recognise, and possible viral changes. The aim is to design candidates that work across variants. Funding and participation describe a research programme. They are not evidence of clinical effectiveness or a promise that a finished vaccine will reach patients by a particular date.
Sources
Vote on this story
Docomo estimates sleep quality from smartphone use
Company claim · Japan
Japan's NTT Docomo announced AI that estimates sleep quality using phone activity, location and personal attributes. It is available through its business health platform, with a city service offering results to residents since October 26.
For people: The tool uses phone and personal data to estimate sleep without requiring a wearable device.
Docomo announced the feature on October 27. AI Watch reports that it uses screen-operation logs, location and attributes such as age to estimate sleep duration, daytime activity and night-time phone use. It also estimates habits associated with a score and can suggest changes. The company says wearable data can improve the estimate, although wearing a device is optional. Musashimurayama's health-points programme began showing results on October 26. This is a company-described estimate based on a medical assessment scale, not evidence that a phone can diagnose a sleep disorder.
Sources
Vote on this story
Nigeria's Dangote refinery adopts industrial AI tools
Company claim · Nigeria
Honeywell's new industrial AI platform is launching at Dangote's refinery in Nigeria, Express Computer reports today. Honeywell estimates more than $60 million in possible annual gross benefits at full throughput; those benefits are not guaranteed.
For people: Refinery engineers will use operating data and AI tools to investigate bottlenecks and equipment problems.
The October 27 report follows Honeywell's October 26 announcement. Forge Plant Cognition combines operating data, engineering knowledge and models of physical processes, using NVIDIA technology and Dell infrastructure. Its intended tasks include finding constraints, examining unusual equipment behaviour and suggesting corrective actions. Honeywell's projected benefits exclude implementation costs and depend on future operating results. The companies also describe support for local workforce development. The refinery's separate capacity-expansion plans should not be credited to AI, and the report does not quantify staffing changes or prove that projected savings have already happened.
Sources
Vote on this story
23 Indian companies showcase AI technology in Dubai
Trusted report · UAE / India
Twenty-three Indian companies are showing AI tools and infrastructure at Dubai AI Festival, India's public broadcaster reports today. Their exhibits include multilingual voice tools, business automation, computing services and cybersecurity at the October 26-27 event.
For people: The showcase connects Indian AI suppliers with potential customers and partners outside India.
The Bharat Pavilion was inaugurated on October 26, with reporting published on October 27. DD News describes exhibits ranging from cloud computing and data-centre services to document processing, finance tools and customer support. Other participants focus on governance, digital identity and cybersecurity. India's Telecom Equipment and Services Export Promotion Council is facilitating the pavilion. The festival is organised by Dubai AI Campus with the UAE's AI ministry. A showcase demonstrates what firms are offering; it does not independently verify their product performance or establish that export contracts have been signed.
Sources
Vote on this story
AI demand adds pressure to the clean-energy transition
Trusted report · Global
Tripling renewable power capacity by 2030 may still fall short of climate needs, IRENA's chief tells CNA today. He identifies AI as one contributor to electricity demand growing faster than earlier forecasts, alongside heating and cooling.
For the planet: AI's electricity needs are part of the wider challenge of expanding clean power while improving energy efficiency.
CNA published the interview on October 27 after speaking with IRENA director-general Francesco La Camera on October 26. The global pledge is to triple renewable capacity to 11.2 terawatts by 2030, but he says higher demand and slow efficiency progress make the task harder. He cites a Chinese data centre using solar, wind and hydropower as an example of renewable supply, while acknowledging that countries face different conditions. His warning concerns the whole energy system. The report does not calculate AI's exact share of the additional demand or say AI alone caused the shortfall.
Sources
Vote on this story