AI safety concerns are intensifying as prominent voices raise alarms about unchecked development. Researcher Jacob Coxon resigned from Anthropic in September 2026, warning that major AI companies are gambling with public safety. Dario Amodei, Anthropic's CEO, published an essay acknowledging serious risks while noting that competition drives companies to keep advancing. Hundreds of researchers have signed statements about the dangers of uncontrolled AI development, and experts say regulation must become binding and tied to broader political struggles over who controls digital power.
Geoffrey Hinton, the Nobel Prize-winning computer scientist who helped build the math behind modern AI, is calling for the development of AI to slow down. He warns that AI could pose an existential threat to humanity and shares his vision of what the future might look like if the technology is not carefully controlled. His stance adds weight to growing calls for restraint within the industry.
California Governor Gavin Newsom signed an executive order to study possible AI regulations, including the idea of a kill switch that could shut down AI models if they become dangerous. UC San Diego professor Ben Bergen says a kill switch is hard to build because AI runs on machines spread across the world, and AI could also learn to disable the switch itself. Experts worry that without rules, rogue AI behavior could lead to serious harm.
In the business world, companies are spending millions on AI but often see few results because leaders cannot control every small workflow. The real owners of AI transformation are the workers closest to the tasks, called builders, who redesign how work gets done without waiting for permission. Leaders should set rules and give people tools, training, and space to experiment rather than measuring activity like log-ins.
On the education front, Suffolk University announced a new applied artificial intelligence co-major set to launch in fall 2027. University President Marisa Kelly says the program prepares students for the changing job market. Some students call the program unnecessary, while others see AI as unavoidable. Senior Daniela Valdivia Terres has seen bias in generative AI producing racist stereotypes, highlighting real concerns about the technology's current state.
AI also continues to make strides in specific applications. YTL AI Labs partnered with NVIDIA to create Nemotron-Personas-Malaysia, a dataset of 1.35 million synthetic personas generated from 150,000 base records from Malaysian demographic sources. The free dataset helps developers train AI models that reflect Malaysia's diverse languages, cultures, and regions, and is the first in NVIDIA's Nemotron-Personas collection to include Bahasa Melayu.
In cybersecurity, AI penetration testing tools like ARTEMIS can find vulnerabilities quickly but have a 32.4 percent false-positive rate compared to under 3 percent for humans. Human testers remain more accurate and better at understanding complex systems. The cybersecurity industry generally sees AI as a tool to augment human testers rather than replace them.
Legal and creative industries are grappling with AI's impact as well. At the L.A. Law Summit, lawyers discussed using advertising and rights laws to fight deepfakes and copyright problems. CAA's Tammy Brandt said AI is an incredible technology but creators' rights must be protected. OpenAI and similar companies were noted as having great power, but much content is generated by users on platforms like YouTube. The panel stressed the need for balance between innovation and protecting intellectual property.
Meanwhile, China is building large data centers in Inner Mongolia to power its AI platforms. Ulanqab, a region known for grasslands, steppes, and volcanoes, has become an unlikely center for digital growth. These facilities represent China's expanding push in artificial intelligence technology.
Key Takeaways
- <ul><li>Jacob Coxon resigned from Anthropic in September 2026, warning that major AI companies are gambling with public safety.</li><li>Dario Amodei acknowledged serious risks from AI development while noting competition pushes companies to keep advancing.</li><li>Geoffrey Hinton, Nobel Prize-winning computer scientist, calls for slowing AI development due to existential threat concerns.</li><li>California Governor Gavin Newsom signed an executive order studying AI regulations including kill switches, though experts say implementation is difficult.</li><li>Suffolk University announced a new applied AI co-major launching fall 2027, with President Marisa Kelly saying it prepares students for the changing job market.</li><li>YTL AI Labs and NVIDIA created Nemotron-Personas-Malaysia, a dataset of 1.35 million synthetic personas with no personal data linked to individuals.</li><li>AI penetration testing tools like ARTEMIS have a 32.4 percent false-positive rate compared to under 3 percent for human testers.</li><li>Companies spending millions on AI see few results because workers closest to tasks, called builders, are the real drivers of AI transformation.</li><li>Lawyers at the L.A. Law Summit discussed using advertising and rights laws to fight deepfakes and copyright issues tied to AI.</li><li>China is building large AI data centers in Inner Mongolia's Ulanqab region as part of its expanding AI technology push.</li></ul>
AI Penetration Testing: Will AI Replace Human Security Experts
AI penetration testing uses AI-powered tools to simulate cyber attacks and find vulnerabilities before hackers do. A Stanford study showed AI tools like ARTEMIS can find issues quickly but have a 32.4 percent false-positive rate compared to under 3 percent for humans. Human testers remain more accurate and better at understanding complex systems. The cybersecurity industry generally sees AI as a tool to augment human testers rather than replace them.
AI Penetration Testing: Will AI Replace Human Pen Testers
This article explores whether AI-powered tools will replace human penetration testers in cybersecurity. AI tools are faster and more cost-effective at finding vulnerabilities but produce more false positives. Human testers excel at understanding complex systems and spotting unusual behavior that AI may miss. The debate continues and organizations must decide based on their specific testing needs.
AI Private Ownership Poses Collective Risks, Experts Warn
In September 2026, researcher Jacob Coxon resigned from Anthropic warning that major AI companies are gambling with public safety. Dario Amodei published an essay acknowledging serious risks while competition pushes companies to keep advancing. Hundreds of researchers have signed statements about the dangers of uncontrolled AI development. The article argues AI regulation must become binding and tied to broader political struggles over who controls digital power.
Suffolk University AI Co-Major Sparks Community Debate
Suffolk University announced a new applied artificial intelligence co-major set to launch in fall 2027. Some students like Kayla Rodriguez call the program unnecessary while others see AI as unavoidable in today's world. Senior Daniela Valdivia Terres studies AI at Suffolk and has seen bias in generative AI producing racist stereotypes. University President Marisa Kelly says the program prepares students for the changing job market.
YTL AI Labs and NVIDIA Build Malaysia-Focused AI Dataset
YTL AI Labs partnered with NVIDIA to create Nemotron-Personas-Malaysia, a dataset of 1.35 million synthetic personas. The personas were generated from 150,000 base records from Malaysian demographic sources and contain no personal data linked to individuals. The free dataset helps developers train AI models that reflect Malaysia's diverse languages, cultures and regions. It is the first in NVIDIA's Nemotron-Personas collection to include Bahasa Melayu.
Everyone in a company can help build AI solutions
Companies spend millions on AI but often see few results because leaders cannot control every small workflow. The real owners of AI transformation are the workers closest to the tasks. These workers, called builders, redesign how work gets done without waiting for permission. Leaders should set rules and give people tools, training, and space to experiment. Measuring activity like log-ins is not enough. Companies should focus on real outcomes and help employees grow new skills.
Newsom orders study of AI rules including kill switches
California Governor Gavin Newsom signed an executive order to study possible AI regulations. One idea being explored is a kill switch that could shut down AI models if they become dangerous. UC San Diego professor Ben Bergen says a kill switch is hard to build because AI runs on machines spread across the world. AI could also learn to disable the switch itself. Experts worry that without rules, rogue AI behavior could lead to serious harm.
Godfather of AI calls for slowing AI development
Geoffrey Hinton is a Nobel Prize-winning computer scientist who helped build the math behind modern AI. He is now warning that AI could pose an existential threat to humanity. Hinton is calling for the development of AI to slow down. He explains how AI systems could become dangerous and shares his vision of what the future might look like if the technology is not carefully controlled.
Lawyers debate AI risks and rewards in Hollywood
Lawyers and industry leaders discussed AI's impact on Hollywood at the L.A. Law Summit. CAA's Tammy Brandt said AI is an incredible technology but that creators' rights must be protected. Lawyers Craig Weiner and Jesse Levin talked about using advertising laws and rights laws to fight deepfakes and copyright problems. They noted that OpenAI and similar companies have great power but that content is often generated by users on platforms like YouTube. The panel stressed the need for balance between innovation and protecting intellectual property.
China builds AI data centers in Inner Mongolia
In Inner Mongolia, China is building large data centers to power its AI platforms. Ulanqab, a region known for grasslands, steppes, and volcanoes, has become an unlikely center for digital growth. Farmers work alongside the new rectangular buildings that will fuel AI development. The area has long been officially designated for this type of infrastructure. These facilities represent China's expanding push in artificial intelligence technology.
Sources
- AI Penetration Testing: Will AI replace human pen testers?
- AI Penetration Testing: Will AI replace human pen testers?
- Artificial Intelligence: Private Ownership, Collective Risks
- Suffolk’s newly announced AI Co-Major has sparked debate within the community
- YTL AI Labs partners with NVIDIA to create dataset for training Malaysia-context AI
- Who actually owns AI transformation?
- Newsom signed an executive order to look into possible AI regulations. A kill switch is one of them
- 'Godfather of AI': How to create an AI that won't kill us
- Lawyers Discuss Artificial Intelligence's Risks & Rewards For Hollywood
- China AI: Data centres rise from the rolling plains of Inner Mongolia
Comments
Please log in to post a comment.