The FAA says its $875 million SMART system did not cause a ground stop issued at about 3:30 p.m. ET on Friday; flights resumed at 8:15 p.m. The tool, developed by Air Space Intelligence and unveiled in September, combines 200 data streams—including weather, routes, traffic and staffing—to recommend options to FAA specialists and controllers. It does not control flights. In an initial Washington-area test, SMART helped reroute 180 flights, saving an average of 10 minutes and about 80 gallons of fuel per flight; one Tampa–Dallas recommendation saved 42 minutes.
The system has drawn calls for safeguards and scrutiny. House Democrats asked the Government Accountability Office to audit SMART, citing questions about safety workers’ input and training and the possibility of false AI information. Aviation safety consultant and former FAA official Mike O’Donnell also called for checks on recommendations and continued human oversight. The FAA says people remain responsible for decisions.
Amazon reported that records collected by its MadPot decoy sensors from Oct. 3 to Nov. 5 showed AI agents searching internet-exposed applications for AI-provider keys, validating working credentials and storing them. Targets included tools such as LiteLLM, One API and Dify, with hundreds of targets described. Amazon assessed with moderate confidence that separate China-linked groups operated the campaigns, citing indicators including Chinese-language prompts and time-zone settings. A separate campaign used deepfakes in attempts to pass identity checks at two cryptocurrency exchanges; a vendor rejected each attempt. Amazon notified potentially affected AWS customers and advised organizations to secure exposed systems and monitor for unusual usage or charges.
Researchers at Sookmyung Women’s University developed X-SPUR, which they say detects cyberattacks on communication networks in connected and autonomous vehicles with 99.87% accuracy, including when vehicles are moving. The system also explains why it classifies activity as an attack. In Japan, Defense Minister Shinjiro Koizumi said the ministry plans to issue updated AI-use guidelines for defense at the end of November. The rules will cover ministry staff and Self-Defense Forces personnel, emphasizing human decision-making and responsibility, data use, and balancing broader adoption with safety.
Asian markets felt pressure from semiconductor and AI-related stock declines: Japan’s Nikkei briefly dropped below ¥60,000 on Nov. 16, reaching the ¥59,800 range. Its morning session closed at ¥60,182.50, down 1.57% from the previous week’s close. Advantest fell more than 10%, while weakness in U.S. semiconductor shares and declines in South Korea and Taiwan weighed on sentiment. Some domestic and defensive shares rose, including Nitori Holdings after strong earnings.
Deloitte Asia Pacific is launching its CEO AI Leadership Report during the 2026 APEC CEO Summit in Shenzhen on Nov. 17–18. It describes an “originality gap” between rapid technological change and businesses’ ability to adapt and innovate, and recommends investment in AI research and development, experimentation, agility and a distinctive value proposition. Separately, an article argues that AI’s benefits should reach more Americans rather than primarily technology leaders, Donald Trump and his allies, but the provided text does not detail its proposed steps.
Other articles focus on AI’s risks and practical use. A Realtors Property Resource survey found 92% of agents surveyed used or planned to use AI, while 63% named accuracy as their top concern. AI-generated property listings can introduce unsupported claims, change disclosures or raise fair-housing concerns, so the article calls for trained human review and clear rules. StoredTech, a managed IT provider in Queensbury, New York, says its Titan-powered AI and automation help technicians resolve problems faster and identify risks sooner while leaving decisions to staff. Its internal data shows response times improved 54% and service tickets fell 25% compared with before adoption.
An independent research nonprofit report says Anthropic’s AI growth is nearing its internal risk thresholds and may have passed the level the company considers safe, as it improves models ahead of a major initial public offering later that week. The provided summary says the report follows repeated calls from CEO Dario Amodei, but ends before explaining them further.
Key Takeaways
- The FAA says its $875 million SMART system was not connected to Friday’s ground stop; it offers recommendations to human decision-makers, not direct flight control.
- SMART combines 200 data streams and has helped reroute 180 flights in testing, saving an average of 10 minutes and about 80 gallons of fuel per flight; one Tampa–Dallas flight saved 42 minutes.
- House Democrats requested a GAO audit of SMART, raising questions about safety workers’ input and training, false AI information and oversight.
- Amazon found AI agents searching exposed AI applications for provider keys; it assessed with moderate confidence that separate China-linked groups operated the campaigns and notified potentially affected AWS customers.
- Amazon also described deepfake attempts to pass identity checks at two cryptocurrency exchanges; a vendor rejected each attempt.
- Sookmyung Women’s University’s X-SPUR is reported to detect cyberattacks on moving connected and autonomous vehicles with 99.87% accuracy and explain its classifications.
- Japan plans updated defense AI guidelines by the end of November, with emphasis on human responsibility, data use and safety.
- The Nikkei briefly fell below ¥60,000 on Nov. 16 as semiconductor and AI-related shares declined; Advantest dropped more than 10%.
- Deloitte Asia Pacific is launching its CEO AI Leadership Report at the 2026 APEC CEO Summit in Shenzhen, calling on businesses to invest in AI R&D, experimentation and agility.
- Anthropic’s growth may be nearing or exceeding its internal safety thresholds, according to an independent nonprofit report; other coverage highlights AI accuracy and oversight concerns in real estate and reports StoredTech’s claimed 54% faster response times and 25% fewer tickets.
FAA says new AI air traffic system did not cause ground stop
The FAA said its $875 million SMART system was not connected to a ground stop after a technical problem on Friday. The agency issued the stop at about 3:30 p.m. ET, and air traffic resumed at 8:15 p.m. House Democrats asked the Government Accountability Office to audit SMART, raising concerns about safety workers’ input and the risk of false AI information. SMART gives recommendations to people rather than controlling flights. Since it began on Sept. 21, the FAA says it has helped reroute 180 flights, saving an average of 10 minutes and almost 80 gallons of fuel per flight.
FAA officials say SMART AI system was not linked to ground stop
Government officials and Air Space Intelligence, the company behind SMART, said the system did not cause the airline ground stop. SMART stands for Strategic Management of Airspace, Routes and Trajectories, and it combines 200 data streams, including weather, flight paths, traffic and staffing information. Experts use it to send recommendations to FAA officials and air traffic controllers, who make the decisions. During an initial test near Washington this fall, the system helped reroute 180 flights, saving an average of 10 minutes and 80 gallons of fuel per flight; one Tampa to Dallas flight saved 42 minutes. House Democrats raised concerns about safety workers’ input, training and the risk of AI producing false information, while the FAA says SMART is designed to assist people, not control flights.
FAA’s SMART AI tool draws safety questions after ground stop
Officials and Air Space Intelligence said the FAA’s $875 million SMART system was not connected to the airline ground stop. Unveiled in September by Transportation Secretary Sean Duffy and FAA Administrator Bryan Bedford, SMART brings together 200 data streams to help experts spot traffic and weather problems. It sends recommendations to people rather than controlling flights. House Democrats and aviation safety consultant Mike O’Donnell have raised concerns about AI reliability and the need to keep human oversight in place.
FAA says SMART AI tool has rerouted 180 flights during testing
The FAA’s SMART system combines 200 data streams to help specialists spot air traffic and weather problems before they arise. During an initial test in the Washington area this fall, it helped reroute 180 flights, saving an average of 10 minutes and 80 gallons of fuel per flight. The FAA said one recommendation cut 42 minutes from a flight between Tampa, Florida, and Dallas. SMART can help find alternate routes, manage airport arrivals and use safe airspace around storms. Officials say it assists human decision makers, but former FAA official Mike O’Donnell called for safeguards to check its recommendations.
Amazon reports AI agents tried to steal AI provider keys
Amazon said its MadPot decoy sensors collected records from Oct. 3 to Nov. 5 showing AI agents searching for AI provider keys. The records describe agents scanning internet accessible applications, taking provider keys, checking them and storing working credentials. Amazon assessed with moderate confidence that the agents used the open source ARTEX framework and that separate China nexus groups operated them, based on Chinese language prompts, time settings and targeting. A second campaign used deepfakes in attempts to pass identity checks at two cryptocurrency exchanges, but one vendor rejected each attempt. Amazon notified AWS customers it thought might be affected and advised organizations to use standard security controls and watch for unusual charges or usage.
Amazon finds AI agents searching for exposed AI app keys
Amazon’s threat intelligence team found decoy records describing a campaign against internet exposed AI apps, including LiteLLM, One API, and Dify. The records say attackers searched websites, code repositories, packages, container images, and other sources for exposed AI provider keys, and covered hundreds of targets. Amazon assessed with moderate confidence that separate China linked groups ran the campaigns, based partly on Simplified Chinese prompts and China time zone settings, and notified AWS customers who might be affected. The records also describe agents trying to bypass identity checks at two cryptocurrency exchanges with deepfakes. Amazon advises treating AI apps and related systems like production systems, assuming exposed secrets will be found quickly and watching for unusual spending or usage.
AI tool detects vehicle cyberattacks with 99.87% accuracy
A team at Sookmyung Women’s University developed X-SPUR, an AI technology designed to detect cyberattacks on communication networks in autonomous and connected vehicles. The system reportedly detects attacks on vehicles in motion with 99.87% accuracy. It also explains why its AI classified activity as a hack. The team includes Professor Jung Seong-hun of the Department of AI Engineering and undergraduate data science student Kim Ji-su.
Japan plans new AI guidelines for defense use by month’s end
Japanese Defense Minister Shinjiro Koizumi said on November 15 that the Defense Ministry plans to release guidelines for defense related AI use at the end of November. The guidelines will update the ministry’s AI policy, which was created in July 2024, and apply to ministry staff and Self-Defense Forces personnel. Koizumi said the rapid pace of AI change makes updates necessary. The guidelines will focus on human decision making and responsibility, using data to strengthen national defense, and balancing wider AI use with safety.
Nikkei slips below ¥60,000 as semiconductor selloff hits Asian markets
Japan’s Nikkei Stock Average briefly fell below ¥60,000 on November 16, reaching the 59,800 range in afternoon trading for the first time since November 10. The morning session ended at ¥60,182.50, down ¥959.85, or 1.57%, from the previous week’s close. Semiconductor and AI related stocks led the decline, with Advantest falling more than 10%, while weakness in U.S. semiconductor shares and sharp drops in South Korea and Taiwan added pressure. Investors bought some domestic and defensive stocks, including Nitori Holdings, which rose on strong earnings.
Deloitte report urges businesses to close the AI originality gap
Deloitte Asia Pacific is launching its CEO AI Leadership Report during the 2026 APEC CEO Summit in Shenzhen, China, on November 17 and 18. The report says businesses face an “originality gap,” the difference between fast technological change and companies’ ability to adapt and innovate. It recommends investing in AI research and development and building a culture of innovation and experimentation. It also urges companies to create a unique value proposition and become more agile and open to change to compete in the AI era.
Building an AI revolution that benefits all Americans
The article argues that Americans should share in the benefits of artificial intelligence, rather than seeing wealth flow mainly to technology leaders, Donald Trump, and his allies. It calls for a broader approach to AI so its gains reach more people. The provided article text ends before explaining the proposed steps.
AI listing copy raises risks of misleading claims and bias
AI helps real estate agents and sellers quickly create listing descriptions and other marketing, but fast production can reduce review and raise legal risks. Even a small wording change may make an unsupported claim, alter a disclosure, or raise fair housing concerns. A 2026 Realtors Property Resource survey found that 92% of agents surveyed used or planned to use AI, while 63% said accuracy was their top concern. The article says human reviewers need training and clear rules, and must check for misleading claims, discrimination, and confidential information. AI can assist with work but cannot replace professional judgment.
Report warns Anthropic AI growth may exceed its safety limits
Anthropic is improving its AI models as it prepares for a major initial public offering later this week. A report from an independent research nonprofit says the company’s growth is nearing its internal risk thresholds and may have passed the level Anthropic considers safe. The report comes after repeated calls from CEO Dario Amodei to......
StoredTech uses AI to speed IT support and spot risks sooner
StoredTech, a managed IT services provider in Queensbury, New York, has expanded its use of AI and automation to improve support, resolve problems faster, and find risks early. Its Titan-powered technology brings together service histories, procedures, client instructions, and diagnostic data, while technicians remain responsible for decisions. The company says automated agents check other clients’ systems when they find a potential risk in one environment. Internal service data shows 54% faster response times and 25% fewer service tickets opened compared with before the technology was adopted. StoredTech plans to keep improving its services using employee expertise and client feedback.
Sources
- FAA's new $875 million AI air traffic system is drawing scrutiny after ground stop
- Questions swirl around FAA's new SMART AI system, though officials say no link to ground stop
- Questions swirl around FAA's new SMART AI system, though officials say no link to ground stop
- Questions swirl around FAA’s new SMART AI system, though officials say no link to ground stop
- Amazon captured records of AI agents trying to steal AI provider keys
- Amazon Uncovers AI Agent Campaign Targeting LiteLLM, Dify, and Other Apps for API Keys
- AI Security Technology Developed to Detect Autonomous Vehicle Hacks With 99.87% Accuracy
- Japan Defense Ministry to Release AI Guidelines at End-Nov.
- Nikkei briefly dips below ¥60,000 as AI, semiconductor selloff spreads across Asian markets
- The ‘originality gap’ and the search for competitive advantage in the AI era
- How to Build an Artificial Intelligence Revolution for All
- Using AI for listings, where marketing ends and misrepresentation begins
- Anthropic AI Acceleration Verges On Risk Thresholds, Group Says
- StoredTech Expands AI-Powered Managed IT Services to Improve Client Support and Proactive Risk Detection
Comments
Please log in to post a comment.