Security researchers from Hacktron AI used Anthropic's Claude to breach OpenAI's ChatGPT systems in under 72 hours, gaining access to an employee account and private source code. OpenAI confirmed the vulnerabilities were patched and paid the team $6,500 through its bug bounty program. The incident underscores growing security concerns across the AI industry.
Meanwhile, AI hiring tools are creating a stuck job market loop. Job seekers now submit an average of 140 applications per vacancy, with 73% using AI to polish their resumes. Yet only 21% of recruiters actually use AI to screen CVs, leaving candidates feeling judged by automated systems they cannot understand.
On the policy front, Montgomery County, Pennsylvania, passed AI guidelines covering transparency, ethics, and bias testing after setting up an AI council about a year ago. Senator Josh Hawley supports making AI companies liable for what their agents do, while Dr. Abby Stylianou of Saint Louis University warns that AI models can produce realistic text, video, and audio.
Google is expanding its AI & Economy Research Program by hiring economists Anu Madgavkar and Daniel Rock as directors. Their team will study how AI affects jobs and the global economy using real-time data, with findings available at ai.google/economy. Huawei also revealed its upgraded Stellar AI Network Solution at HUAWEI CONNECT 2026 in Shanghai, featuring new switches that cut costs and reduce delays along with strong threat detection measures.
Data center costs remain a pressing issue, with building and powering facilities running between $1 million and $5 million per megawatt according to the Natural Resources Defense Council. Leaders of major AI companies have also warned that AI could help create dangerous bioweapons, citing a 2022 study where an AI model easily generated thousands of molecules usable as chemical weapons.
Key Takeaways
- Hacktron AI used Anthropic's Claude to breach OpenAI's ChatGPT in under 72 hours, accessing an employee account and private source code
- OpenAI patched the vulnerabilities and paid researchers $6,500 through its bug bounty program
- Job seekers face an average of 140 applications per vacancy, with 73% using AI to polish resumes while only 21% of recruiters use AI screening
- Montgomery County, Pa. passed AI guidelines covering transparency, ethics, and bias testing
- Google hired economists Anu Madgavkar and Daniel Rock to study AI's impact on jobs and the global economy
- Data center costs range from $1 million to $5 million per megawatt according to the Natural Resources Defense Council
- Huawei unveiled its upgraded Stellar AI Network Solution at HUAWEI CONNECT 2026 in Shanghai
- AI company leaders warn that AI could help create dangerous bioweapons, referencing a 2022 study on chemical weapon molecule generation
- Senator Josh Hawley supports making AI companies liable for what their AI agents do
- Dr. Abby Stylianou of Saint Louis University warns AI models can produce realistic text, video, and audio
Security researchers used Claude AI to hack ChatGPT
Researchers from Hacktron AI used Anthropic's Claude AI platform to hack into OpenAI's ChatGPT. They gained access to an employee's ChatGPT account and retrieved source code storage data. The entire breach happened in less than 72 hours. The team disclosed the vulnerability in a blog post on Sunday.
Cyber researchers breached OpenAI using Anthropic's Claude
A small cyber security group used Anthropic's Claude AI model to break into OpenAI's systems. They accessed an employee's ChatGPT account and read private software information. The researchers were paid for the work as part of a security research project. The incident highlights risks in AI security and the need for better safeguards.
AI hiring tools create a stuck job market loop
Job seekers face an average of 140 applications per vacancy, with 73% using AI to polish resumes and cover letters. Many candidates believe AI screens CVs, but only 21% of recruiters actually use that technology. AI interviewers and automated hiring tools leave candidates feeling judged and misunderstood. Experts warn that similar AI models across employers may favor the same candidates repeatedly.
Experts urge guardrails as AI concerns grow in St. Louis
Dr. Abby Stylianou of Saint Louis University warns that AI models can produce realistic text, video and audio. Senator Josh Hawley supports making AI companies liable in court for what their AI agents do. Some leading AI researchers have quit over safety concerns. Silicon Valley executives have called for slower AI development, though motivations may vary.
The debate over who pays for AI data center power costs
Building and powering data centers costs between $1 million and $5 million per megawatt according to the Natural Resources Defense Council. These costs burden both companies and the electrical grid. Electricity production for data centers contributes significantly to greenhouse gas emissions. Axios Live explores solutions including renewable energy and more efficient data center designs.
Should Congress limit how AI develops
The fast growth of AI has caused a debate about whether Congress should set limits on its development. Some experts say rules are needed to stop AI from being used for bad purposes. Others worry that limits would slow down new ideas and progress. The public is being asked to share their opinions on the matter through a poll.
Huawei reveals upgraded Stellar AI Network
At HUAWEI CONNECT 2026 in Shanghai, Huawei introduced its upgraded Stellar AI Network Solution. The new system focuses on secure and intelligent connectivity for the AI era. It includes Stellar AI Fabric 2.0 with new switches that cut costs and reduce delays. It also features strong security measures that detect threats like hashrate theft and prompt injection with high accuracy.
AI bioweapons risk warns biotech industry
Leaders of major AI companies have warned that AI could help create dangerous bioweapons. In 2022, researchers found that an AI model easily generated thousands of molecules that could be used as chemical weapons. Experts say AI tools are now widely available and could be misused for harmful biology. Some scientists urge stronger safeguards and better monitoring while others believe current risks are overstated.
Google adds new experts to AI research team
Google is expanding its AI & Economy Research Program by hiring leading economists. New directors Anu Madgavkar and Daniel Rock will study how AI affects jobs and the global economy. The team will use real-time data to help workers and policymakers prepare for changes brought by AI. Research findings and data are available at ai.google/economy.
Hackers break into OpenAI systems
A group of cybersecurity researchers called Hacktron breached OpenAI earlier this year in a 72-hour operation. OpenAI confirmed the vulnerabilities have been patched and paid the researchers $6,500 through its bug bounty program. Experts say the breach shows how important constant security vigilance is for AI companies. The incident adds to growing concerns about AI safety and security worldwide.
Montgomery County, Pa., takes steps to oversee AI use
Montgomery County Commissioners held a meeting to discuss artificial intelligence and public policy. Commissioner Neil Makhija said the county set up an AI council about a year ago. The council produced guidelines that were passed at a recent meeting. The guidelines cover transparency, ethics, testing for bias, and other areas. Commissioner Tom DiBello expressed skepticism about AI fears from tech CEOs but supported putting guidelines in place.
Sources
- AI security experts say they used Claude to hack ChatGPT
- Researchers used Claude to hack OpenAI
- AI is reading AI and hiring is stuck
- Navigating concerns over artificial intelligence
- Axios Live: The fight over who pays for AI's power boom
- Do you think Congress should impose some limits on the development of artificial intelligence?
- Huawei Unveils Upgraded Stellar AI Network to Power the Agentic World with Secure, Intelligent Connectivity
- The specter of AI-enabled bioweapons is a wake-up call for biotech
- New experts join Google’s AI & Economy team
- Hackers breached OpenAI, adding to fever pitch of security and safety concerns
- Montgomery County, Pa., Has ‘Taken a Number of Steps’ on AI
Comments
Please log in to post a comment.