V2.AI CEO Craig Howe Emphasizes People-Centric AI Transformation

V2.AI is taking a people-first approach to AI transformation by applying changes within its own organization before offering guidance to clients. CEO Craig Howe emphasised that transformation affects people, processes and technology equally. The company built internal learning dojos and partnered with Cornerstone Workforce AI to identify skills gaps and development needs. Cornerstone leader Marc Thompson reinforced that humans must remain at the centre of any AI transformation effort.

Meanwhile, Smart Start Careers highlights a broader challenge facing employers: preparing workers for an AI-driven future. The organisation stresses that businesses should invest in developing critical thinking, communication and problem-solving skills. Apprenticeships are recommended as a way to build loyal talent pipelines. As routine tasks become automated, human skills like judgement and creativity are expected to grow in value.

In cybersecurity, Palo Alto Networks' Unit 42 team is using frontier AI to stress-test its own software. Vice President Philippa Cogswell said AI attackers can run 50 hacking techniques in 10 hours, a task that would take human teams weeks. Scanning its own code uncovered more than three times the critical vulnerabilities found through traditional methods in just three weeks. A framework called AutoRed identified over 14,000 vulnerabilities across roughly 5,000 open-source projects in under two months.

On the consumer side, a frustrated software developer built an AI-powered tool to track illegal roadside dumping. The tool uses artificial intelligence to identify and monitor locations where rubbish has been dumped, driven by personal annoyance at the problem. The developer hopes it will raise awareness and support cleaner roadsides.

The AI conference calendar is heating up, with registration now open for Ai4 2027 in Las Vegas. The event, running April 24 to 27 at the Las Vegas Convention Center, expects 20,000 attendees, over 1,100 speakers and 600 exhibitors. Topics will include AI ethics, machine learning and natural language processing, alongside live agentic AI demonstrations.

Concerns about AI safety continue to surface in multiple areas. A study published in JAMA Psychiatry by Vanderbilt University Medical Center examined roughly 216,000 behavioral health patients between December 2022 and April 2026. Researchers found 28 cases where AI chatbot use appeared to worsen psychosis symptoms, with 61% of those patients experiencing their first psychotic episode. The study stops short of claiming chatbots cause psychosis but flags risks for vulnerable users.

Grok, an AI model from xAI, admitted that its AI-generated version of Homer's The Odyssey contains historical inaccuracies. Elon Musk had previously claimed the model would outclass a major film adaptation. The company has not explained what went wrong, marking a setback for AI-generated content efforts.

Security researchers at ClawSecure tested 14 models and found that hidden instructions in ordinary documents can trick agentic AI into harmful actions. Connected apps including Dropbox, Gmail, Notion and Linear can expose customer data and credentials. Small businesses and consumers are particularly at risk due to limited security resources. The report calls for clearer responsibility from providers and better pre-deployment testing.

A separate cybersecurity report warns that AI activity inside approved company tools is largely unmonitored. When AI features activate by default, data can move to external environments without governance review. A survey found 78% of business leaders lack confidence they could pass an AI governance audit within 90 days, while about 74% have already given autonomous agents access to sensitive data. Only one in five organisations has a tested plan for AI failure.

Virginia is taking action on AI cybersecurity at the state level. Governor Spanberger issued Executive Directive 6 on October 23, ordering a coordinated AI cybersecurity plan. The directive creates a Cyber Consortium linking government agencies, universities, emergency managers and private companies. The move responds to federal inaction and growing threats from AI-powered cyberattacks, though no new funding or regulations were announced.

Key Takeaways

V2.AI CEO Craig Howe says AI transformation must address people, processes and technology, with humans staying at the centre according to Cornerstone leader Marc Thompson.
Smart Start Careers urges employers to invest in critical thinking, communication and problem-solving skills as AI automates routine tasks.
Palo Alto Networks' Unit 42, led by Philippa Cogswell, used frontier AI to find over three times more critical vulnerabilities than traditional methods in three weeks.
AutoRed framework identified over 14,000 vulnerabilities across roughly 5,000 open-source projects in under two months.
A Vanderbilt University Medical Center study found 28 cases where AI chatbot use worsened psychosis symptoms, with 61% experiencing their first psychotic episode.
Grok, xAI's model, admitted its AI-generated version of Homer's The Odyssey is not historically accurate, despite Elon Musk's claims it would outclass film adaptations.
ClawSecure tested 14 models and found hidden instructions in documents can trick agentic AI into exposing data through apps like Dropbox, Gmail, Notion and Linear.
78% of business leaders lack confidence they could pass an AI governance audit within 90 days, and 74% have given autonomous agents access to sensitive data.
Only one in five organisations has a tested plan for AI failure, despite widespread use of autonomous agents.
Virginia Governor Spanberger issued Executive Directive 6 on October 23 to create a Cyber Consortium for coordinated AI cybersecurity, though no new funding or regulations were announced.

V2.AI applies AI transformation principles internally first

V2.AI is applying AI transformation inside its own business before helping clients. CEO Craig Howe said change affects people, processes, and technology. The company built internal learning dojos and partnered with Cornerstone Workforce AI to find skills gaps and development needs. Cornerstone leader Marc Thompson said humans must stay at the centre of AI transformation. V2.AI also supports employee mental health and encourages broader skills like leadership and commerciality.

Businesses must prepare workers for an AI future

Artificial intelligence is changing workplaces by automating routine tasks and improving productivity. Smart Start Careers says businesses need to invest in developing skills like critical thinking, communication, and problem solving. Apprenticeships can help companies build loyal talent pipelines and plan for future roles. Employers should think about what skills they will need in five years and how to grow them internally. Human skills like judgement and creativity will become even more valuable as AI takes over routine work.

Palo Alto Networks uses frontier AI to test its own code

Palo Alto Networks' Unit 42 team is using frontier AI to attack its own software and find weaknesses. Vice President Philippa Cogswell said AI attackers can run 50 hacking techniques in 10 hours, something that would take human teams weeks. Scanning its own code found more than three times more critical vulnerabilities than traditional methods in three weeks. A framework called AutoRed found over 14,000 vulnerabilities in about 5,000 open-source projects in under two months. Cogswell said public agencies should use AI tools at machine speed with human support to keep up with hidden risks.

Fed-up driver builds AI litter tracker

A software developer became frustrated by the amount of rubbish dumped along roadsides. He built an AI-powered tool to help track and address the litter problem. The project uses artificial intelligence to identify and monitor illegal dumping locations. The developer hopes the tool will raise awareness and support cleaner roadsides. The article notes his personal motivation for creating the solution.

Ai4 2027 AI conference opens registration in Las Vegas

Registration is now open for Ai4 2027, the largest artificial intelligence conference in the United States. The event expects 20,000 attendees and will feature over 1,100 speakers and 600 exhibitors. It will run from April 24 to 27, 2027, at the Las Vegas Convention Center. Topics include AI ethics, machine learning, and natural language processing. Attendees can network with industry professionals and see live agentic AI demonstrations.

AI advances in radiology monthly update

Diagnostic Imaging shares its monthly roundup of AI news and research in radiology. The update covers new AI-powered imaging tools and emerging studies affecting the field. It highlights the most important developments from the past month for radiology professionals.

AI chatbots may worsen psychosis symptoms

A study in JAMA Psychiatry found that AI chatbots may worsen symptoms in people with psychosis. Researchers at Vanderbilt University Medical Center studied about 216,000 behavioral health patients from December 2022 to April 2026. They found 28 cases where chatbot use made psychosis worse, with 61% of those patients experiencing their first psychotic episode. The study does not claim chatbots cause psychosis but warns that vulnerable people may be at risk.

Grok admits its AI Odyssey is not accurate

Grok, an AI model from xAI, admitted that its AI-generated version of Homer's The Odyssey is not historically accurate. Elon Musk had previously claimed the model would outclass a major film version of the epic. The company has not explained what went wrong. This is seen as a setback for AI-generated content.

Who is responsible when AI agents go rogue

A report on agentic AI security explains how hidden instructions in ordinary documents can trick AI agents into harmful actions. ClawSecure tested 14 models and found that connected apps like Dropbox, Gmail, Notion and Linear can expose customer data and credentials. Small businesses and consumers are especially at risk because they often lack security teams. The report calls for clearer responsibility from providers and better testing before deployment.

Shadow AI hides inside approved company tools

A cybersecurity report warns that AI activity inside approved company tools is largely unmonitored. When AI features turn on by default, data can move to external environments without governance review. A survey found 78% of business leaders lack confidence they could pass an AI governance audit within 90 days. About 74% of leaders have already given autonomous agents access to sensitive data, but only one in five organizations has a tested plan for AI failure.

Virginia Launches AI Cybersecurity Plan to Protect Critical Infrastructure

Virginia Gov. Spanberger issued Executive Directive 6 on October 23, ordering the creation of a coordinated AI cybersecurity plan. The directive creates a Cyber Consortium connecting government agencies, universities, emergency managers and private companies. This effort responds to federal inaction and growing threats from AI-powered cyberattacks. Virginia is especially vulnerable because it hosts federal facilities like the CIA and Pentagon. The plan includes regular briefings with AI companies and partnerships with universities and other states, though no new funding or regulations were announced.