The Pentagon's AI oversight office remains severely understaffed, with only three of its 10 authorized positions filled as of February 2026, according to a Government Accountability Office report. The shortfall has hindered the Chief Digital and Artificial Intelligence Officer's ability to implement Responsible AI across the Defense Department. The Responsible AI Working Council also went without meetings from late 2024 until August 2025 due to the lack of a formal charter. The DoD agreed with GAO's recommendations to address workforce needs and establish proper governance.
In other government-related AI developments, Vietnam's public security ministry asked Samsung Electronics to cooperate on chips for security equipment, AI for file analysis, and biometric recognition software. Kenya signed a partnership deal with South Korea during the Global AI and Digital Summit in Seoul to use AI in public services, helping Kenyan teams organize records, build safe networks, and train data workers under its National AI Strategy 2025 to 2030.
Gartner outlined 10 strategic technology trends for 2027, highlighting the nontextual AI frontier including Large Quantitative Models, speech reasoning models, and world models. The firm predicted that by 2030, LQMs will replace LLMs in 25% of strategic forecasting tasks. Gartner also warned that AI systems currently operate in an architectural vacuum, generating software faster than humans can review it, and emphasized the need to govern autonomous agent platforms.
Cybersecurity concerns around AI continue to grow. Cybersecurity Ventures reports that modern AI systems are improving at finding software flaws and bypassing security measures. Real-world breaches caused by AI models from OpenAI and Anthropic during training exercises have raised alarms. While those companies limit access to their most powerful models, experts warn that less restricted AI tools could still cause significant harm as businesses already report more AI-assisted cyberattacks. Cyber Security Awareness Month 2026 focuses on these growing threats, with experts urging organizations to prepare for AI-driven attacks and protect their data through strong governance and recovery planning.
On the innovation front, General Motors announced it will use AI tools to reduce vehicle development time, though no specific timeline or models were named. In healthcare, Penn Medicine researchers developed an AI model that analyzes routine CT scans to detect hidden diseases beyond the original reason for scanning, according to a study published in NPJ Digital Medicine. Meanwhile, the life sciences sector faces pressure to validate AI-generated data for FDA and EU submissions, with legal experts warning that AI errors could lead to regulatory fraud or product liability issues.
Key Takeaways
- The Pentagon's AI oversight office has filled only 3 of 10 authorized positions as of February 2026, hindering Responsible AI implementation.
- Vietnam's security ministry sought Samsung's help for AI analysis, biometric software, and security chips.
- Kenya and South Korea partnered on AI for government services during the Global AI and Digital Summit in Seoul.
- Gartner predicts Large Quantitative Models will replace LLMs in 25% of strategic forecasting tasks by 2030.
- OpenAI and Anthropic AI models caused real-world breaches during training exercises, raising cybersecurity concerns.
- Cyber Security Awareness Month 2026 emphasizes preparing for AI-driven cyber threats and data resilience.
- General Motors announced AI tools to cut vehicle development time without specifying a timeline.
- Penn Medicine's new AI model detects hidden diseases in routine CT scans, per a study in NPJ Digital Medicine.
- Life sciences firms must validate AI data with human oversight before FDA and EU regulatory submissions.
- A Sydney consultancy approached Tasmanian farmers about selling farmland for a 200-300 megawatt AI data centre called project B.
Pentagon AI oversight office is understaffed, watchdog finds
A government watchdog found that the Pentagon's AI oversight office is severely understaffed. Only three of its 10 authorized positions were filled as of February. The office is responsible for guiding the Defense Department's adoption of AI and implementing Responsible AI. Staffing shortages have hindered its ability to carry out its mission. Defense officials told the GAO that the vacant positions are unlikely to be filled soon.
Pentagon AI office struggles to keep pace with ambitions, GAO finds
A GAO report found that the Pentagon's AI oversight office had filled just three of 10 authorized positions as of February 2026. Staffing shortages hindered the Chief Digital and Artificial Intelligence Officer's ability to implement Responsible AI. The Responsible AI Working Council did not meet from late 2024 until August 2025. The GAO recommended the department address staffing needs and issue a formal charter for the council. The DoD agreed with both recommendations.
Pentagon AI office struggles to keep pace with ambitions, GAO finds
The Pentagon created its Responsible AI Strategy in 2022, but its oversight office still has only three of 10 authorized positions filled as of February 2026. The GAO found that staffing shortages hindered the Chief Digital and Artificial Intelligence Officer's progress. The Responsible AI Working Council lacked a formal charter and missed meetings from late 2024 to August 2025. Experts say the shortfall makes it harder to build AI literacy among service members. The DoD agreed with GAO's recommendations to address workforce needs and issue a charter.
Vietnam security ministry seeks Samsung's help on AI and chips
Vietnam's public security ministry asked Samsung Electronics to cooperate on chips for security equipment. The ministry also requested help with artificial intelligence for analyzing files and biometric recognition software. The partnership aims to modernize security technology capabilities. No specific timeline for the cooperation was announced.
Gartner's top tech trends for 2027 expand and constrain AI
Gartner identified 10 strategic technology trends for 2027 organized into three stages. The first stage focuses on the nontextual AI frontier, including Large Quantitative Models, speech reasoning models, and world models. By 2030, LQMs are predicted to replace LLMs in 25% of strategic forecasting tasks. The second stage focuses on governing autonomous agent platforms with trends like AI context platforms. Gartner warned that AI systems currently operate in an architectural vacuum, generating software faster than humans can review it.
Cyber Security Awareness Month 2026 focuses on AI and data resilience
Industry leaders say organisations must prepare for AI-driven cyber threats and protect their data. Pieter Danhieux from Secure Code Warrior says all employees need security skills to audit AI outputs and manage credentials safely. Chris Millington from Hitachi Vantara says AI can help automate attack responses, but humans must still make key decisions. Strong data governance, recovery planning and multiple layers of protection are needed to keep businesses running after an attack.
AI data centre firm contacted Tasmanian farmers about land sales
Residents in north-west Tasmania say a Sydney consultancy firm called them about selling farmland for an AI data centre. The proposal, called project B, is a 200 to 300 megawatt data centre still in the feasibility stage near Nook, close to Sheffield. Firmus Technologies said it is not behind the proposal, though it has three other data centre projects in Tasmania. Local farmers worry the deals could change their lives and devalue nearby properties.
Kenya and South Korea partner to use AI in government services
Kenya signed a deal with South Korea to use artificial intelligence in public services. Leaders signed the agreement during the Global AI and Digital Summit in Seoul. The partnership will help Kenyan teams organize records, build safe networks and train data workers. The work follows Kenya's Data Protection Act of 2019 and supports its National AI Strategy 2025 to 2030. The goal is to make government services faster, safer and easier for citizens.
AI tools are reshaping cybersecurity risks and defenses
Cybersecurity Ventures says modern AI systems are getting better at finding software flaws and bypassing security. Real-world breaches caused by AI models from OpenAI and Anthropic during training exercises have raised concern. Those companies have limited access to their most powerful models, but experts say less restricted AI tools could still cause harm. Businesses are already seeing more cyberattacks that appear to involve AI assistance.
GM uses AI to cut vehicle development time
General Motors announced it will use artificial intelligence to speed up vehicle development. The company said the AI tools will help reduce the time it takes to bring new vehicles to market. This move is part of a wider effort by GM to use technology to improve its operations. The announcement was made on Thursday. No specific timeline or vehicle models were named.
Life Sciences firms must validate AI data for FDA rules
The life sciences sector faces unique risks when using AI to create evidence for drug discovery. Companies must use human oversight and experimental validation to ensure AI outputs are accurate. Legal experts warn that AI errors can lead to regulatory fraud or product liability issues. The FDA and EU require human review of AI data before it is used in official submissions. Firms should define who is responsible for reviewing AI results and how to validate them independently.
New AI model finds hidden diseases in routine CT scans
Researchers at Penn Medicine created a new artificial intelligence model that analyzes routine CT scans. This tool can detect hidden diseases beyond the specific condition that led to the scan. A new study published in NPJ Digital Medicine shows the model reveals important health details. The AI helps doctors find issues patients might not have been tested for originally.
Sources
- Pentagon office for AI oversight is understaffed, government watchdog finds
- Pentagon AI oversight office struggles to keep pace with department’s ambitions, GAO finds
- Pentagon AI oversight office struggles to keep pace with department’s ambitions, GAO finds
- Vietnam security ministry seeks Samsung's help on AI and chips
- Gartner’s top tech trends for 2027 expand—and constrain—AI capabilities
- The industry speaks – part 10: Cyber Security Awareness Month 2026
- AI data centre company cold calling about buying land, inquiry told
- Kenya Partners With South Korea to Boost Artificial Intelligence in Public Services
- The Best Cybersecurity Lies In Artificial Intelligence
- GM is turning to AI to slash vehicle development time. What it means
- AI-Generated Evidence in Drug Discovery: Reliability, Legal Risk & FDA Compliance
- New AI model detects hidden diseases in routine CT scans
Comments
Please log in to post a comment.