#AI

All articles tagged with "AI"

New

News

Canva launches Craft, an image model built to plan designs first

Key Takeaways

Canva Craft uses a two-stage system that plans a design before producing an image.
The model supports more than 30 aspect ratios and image generation up to 2K resolution.
Canva says Craft was trained on millions of human-made templates from its content library.

DT Editorial Team·Oct 23, 2026·via theverge.com

RoboBusiness Rebrands as Robotics Summit & Expo West

Key Takeaways

The renamed event is scheduled to debut in Santa Clara in October 2027.
Its program will focus on scale-up issues such as fleet management, service and go-to-market strategy.
The organizer says the shift reflects a market moving beyond prototype-stage robotics.

DT Editorial Team·Oct 21, 2026·via therobotreport.com

More in AI & Robotics→

Reservoir Farms during its grand opening.

AI & Robotics

Washington incubator puts agricultural robots into working fields

Key Takeaways

Reservoir and Washington partners launched an incubator for agricultural-robotics startups working with growers and researchers.
The program targets difficult-to-automate crops including apples, cherries, pears, blueberries, wine grapes and hops.
It extends Reservoir’s on-farm innovation-center strategy, which began with a specialty-crop hub in California’s Salinas Valley.

DT Editorial Team·Oct 21, 2026·via therobotreport.com

More in AI & Robotics→

News

Google Opens Prompt-Built Personal Android Apps to Users

Key Takeaways

Google AI Studio now supports prompt-built personal Android apps for standard Google accounts.
Apps can be privately installed on the creator’s own device and are not publicly listed in Google Play.
Google plans sharing with close family and friends, with a potential path to a professional developer account for apps that gain traction.

DT Editorial Team·Oct 15, 2026·via zdnet.com

AI Victim Impact Video Leads Arizona Court to Order Resentencing

Key Takeaways

Arizona appellate court ordered resentencing because an AI-generated victim impact video carried 'undue emotional weight.'
Gabriel Horcasitas was convicted of manslaughter in the 2021 shooting death of Christopher Pelkey.
Stacey Wales scripted an AI likeness of her brother to forgive the killer; she told 404 Media that swaying the judge was the goal.
The ruling keeps the manslaughter conviction but requires the judge to reconsider the prison term.

DT Editorial Team·Oct 7, 2026·via 404media.co

More in AI & Robotics→

Chatbots as oracles: How AI is giving new life to the ancient practice of seeking answers from inscrutable sources

AI & Robotics

Chatbots as Oracles: How AI Revives Ancient Answer-Seeking

Key Takeaways

A 2026 Pew report found that 10% of Americans use chatbots for emotional support or companionship.
A study of 1.1 million ChatGPT conversations showed 49% of messages involved seeking information or advice.
Ask Delphi, an AI ethics tool, received 3 million queries in weeks, raising concerns about users treating it as a moral authority.
Researchers Webb Keane and Ambuj Tewari draw parallels between ancient oracles like Delphi and modern conversational AI.

DT Editorial Team·Sep 29, 2026·via phys.org

More in AI & Robotics→

A Dragon capsule approaching the ISS. Credit - NASA / Chris Cassidy

Space

Stanford's 'World Model' Teaches Spacecraft to Dream Their Way to Docking

Key Takeaways

Docking with the ISS requires matching speeds near 28,000 km/h, where accelerating forward causes an upward drift and there is no air friction to slow a craft.
Traditional docking uses Guidance, Navigation and Control algorithms with Extended Kalman Filters, but Kalman filters struggle with high-speed video and computer vision is fragile in glare or shadow.
Reinforcement learning can fail when rules change, such as a different ISS docking port orientation.
Stanford's Out-of-this-World Model (OWM) uses a 'World Model' to run mental simulations, inspired by how a baseball outfielder predicts a fly ball without solving equations.

DT Editorial Team·Sep 24, 2026·via universetoday.com

Bad Bot Traffic Surges 124%, Growing Nine Times Faster Than Human Web Traffic

Key Takeaways

Bad bot traffic grew 124% from July 2025 to June 2026, nine times faster than human traffic.
AI traffic increased more than 82%, with scraping up 185% and scalping bots nearly tripling.
AI bots target login pages at eight times the rate seen in the previous year's report.
DataDome detected 52.7 billion crawler requests from AI agents, with Meta (46%) and OpenAI (35%) dominating.

DT Editorial Team·Sep 23, 2026·via fastcompany.com

How AI and Dynamic Sharing Could Unclog Crowded Airwaves

Key Takeaways

Radio spectrum sits at the low end of the electromagnetic spectrum and carries cellphones, TV, radar, GPS, and satellite sensing signals.
Governments historically assigned spectrum through exclusive licenses that reserved specific frequencies for single users.
Long-term measurements of actual use found that large swaths of licensed spectrum sit idle much of the time, such as military radar bands active only in certain places and on certain days.
Dynamic spectrum sharing would let multiple users occupy the same frequencies under clear rules that prevent interference.

DT Editorial Team·Sep 13, 2026·via fastcompany.com

More in AI & Robotics→

Apple pitches its new phones, watches, and AirPods as vehicles for AI

Innovation

Apple Positions New Phones, Watches, AirPods as Personal AI Hubs

Key Takeaways

Apple CEO John Ternus positions new devices as personal AI hubs.
A20 Pro chip with dual 16-core Neural Engine doubles on-device AI compute.
New Siri AI is delayed but will ship in beta later this quarter.
Duo foldable offers dual-screen AI conversations with Siri.

DT Editorial Team·Sep 10, 2026·via fastcompany.com

AI Can't Fix the Emergency Room — And That's the Point

Key Takeaways

AI tools like scribes and triage algorithms are being aggressively marketed as ER fixes, but they only address superficial symptoms.
Emergency departments are overwhelmed due to boarding, staffing shortages, and social determinants of health that no technology can solve.
Brittany Trang's AI Prognosis column argues the ER's problems mirror broader systemic failures in healthcare.
Solving 'dumb problems' like interoperability and workflow is necessary but insufficient.

DT Editorial Team·Sep 9, 2026·via statnews.com

DeepSeek plans massive Huawei AI chip cluster in Inner Mongolia

Key Takeaways

DeepSeek reportedly plans an Inner Mongolia inference cluster using at least 160,000 Huawei Ascend-950DT chips.
The buildout would be the largest known Huawei chip cluster if completed.
The project underscores both China’s AI self-reliance push and its ongoing memory and production bottlenecks.

DT Editorial Team·Sep 4, 2026·via the-decoder.com

More in AI & Robotics→

Practice helps the brain separate tasks and improve multitasking ability

Health

Practice appears to rewire multitasking in the brain

Key Takeaways

Researchers tracked the same neurons over weeks as mice learned a dual-task challenge.
The brain initially showed interference from shared neural resources and coordination demands.
With practice, more neurons became task-specific and task representations separated.
The authors said the findings could help inform AI systems trained for multitasking.

DT Editorial Team·Sep 4, 2026·via medicalxpress.com

DOJ backs fair use defense for AI training in NYT case

Key Takeaways

The U.S. Department of Justice argued that AI training on copyrighted text can qualify as fair use.
The filing draws a distinction between copying during training and the outputs models generate.
The position could shape a landmark lawsuit involving The New York Times, OpenAI, and Microsoft.

DT Editorial Team·Sep 3, 2026·via the-decoder.com

More in AI & Robotics→

Tim Cook handed $47m package for new role as executive chair of Apple

Culture

Apple’s leadership handoff puts pay, AI and succession in focus

Key Takeaways

Apple has transferred the CEO role from Tim Cook to John Ternus while keeping Cook as executive chair.
Compensation details underscore a dual strategy of continuity for Cook and performance pressure for Ternus.
The leadership change arrives as Apple faces rising expectations around AI integration and new hardware launches.

DT Editorial Team·Sep 2, 2026·via theguardian.com

AI age checks become a frontline policy issue for social platforms

Key Takeaways

Meta’s $18 billion settlement puts age assurance at the center of child-safety enforcement.
Platforms including Meta and Google are using AI to estimate whether users are adults or minors.
AI-based age estimation may reduce friction versus ID checks, but it raises privacy and accuracy concerns.

DT Editorial Team·Sep 2, 2026·via fastcompany.com

John Ternus Takes Over as Apple CEO, Opening a New Era

Key Takeaways

John Ternus officially became Apple CEO on September 1, 2026.
Tim Cook moved to executive chairman on Apple’s board.
The transition comes as Apple faces AI competition, component shortages, and executive turnover.
Apple’s September 9 hardware event will be the first major public test of the new era.

DT Editorial Team·Sep 1, 2026·via theverge.com

Skild AI Says Its S1 Model Lets Robots Learn From One Video

Key Takeaways

Skild AI introduced S1 as its flagship robot foundation model.
The company says S1 can learn new tasks from a single human demonstration video.
Skild AI trains on teleoperation, human video, simulation, and glove-capture data.

DT Editorial Team·Sep 1, 2026·via therobotreport.com

More in AI & Robotics→

Japan seeks new extended-range missiles in record $55B defense budget request - Breaking Defense

Military

Japan targets longer-range strike options in record defense budget request

Key Takeaways

Japan requested a record 8.89 trillion yen defense budget for the next fiscal year.
The filing highlights interest in anti-radar missiles, low-cost cruise missiles and submarine-launched systems.
Drones and AI remain central pillars in the ministry’s procurement direction.

DT Editorial Team·Aug 31, 2026·via breakingdefense.com

Google’s WikiSkill gives AI agents a memory of failure

Key Takeaways

WikiSkill pairs AI agents with a persistent wiki-like knowledge base that records failures and successes.
The framework separates raw execution traces, accumulated knowledge, and active procedural skills.
A validation gate tests proposed skill changes and can roll them back if performance drops.
The approach does not retrain the model; it improves future runs through documented, reusable instructions.

DT Editorial Team·Aug 30, 2026·via the-decoder.com

More in AI & Robotics→