AI researcher Jacob Coxon recently left Anthropic, citing fears that the current race to build super intelligent agents could threaten human existence by the end of the decade. He argues that major companies are moving too fast and that existing safety plans are insufficient to prevent potential disasters.
Despite these warnings, the industry continues to expand rapidly. The physical AI market is projected to grow at an annual rate of more than 30 percent over the next decade, with forecasts predicting over 1.3 billion AI-powered robots and autonomous systems operating worldwide by 2035.
Corporate adoption is also accelerating, with the global e-learning market expected to reach $335 billion by 2030. To meet this demand, companies are turning to AI tools like Articulate's Frontline to help employees create training content faster without relying solely on specialist teams.
Infrastructure challenges are shifting focus from hardware to data movement. While GPUs remain important, networking and storage now play a bigger role in performance. Gartner predicts global AI spending will hit about $2.59 trillion, with nearly half dedicated to infrastructure.
Partnerships are streamlining access to these technologies. Talkdesk and Microsoft have formed a new alliance allowing businesses to buy Talkdesk Customer Experience Automation tools directly through the Microsoft Azure Marketplace, making it easier to modernize contact centers.
Meanwhile, institutions like Stony Brook University are deploying physical AI, having received $298,000 to build Rubo, a humanoid robot designed to assist students on campuses with navigation and multilingual support.
However, significant risks remain. A new report warns that AI automation could cost up to 800 million jobs worldwide by 2030, prompting calls for governments and schools to invest heavily in education and green energy policies to create new opportunities.
Technical reliability is another hurdle. IBM researchers found a 24-point gap in consistency when AI agents run the same task multiple times, often failing to complete all required steps. They are developing new systems using memory and guidelines to improve reliability.
In the financial sector, innovation continues at the Global Fintech Fest in Mumbai. Companies like NPCI launched an Agentic AI platform and UPI Tap & Pay, while others introduced features for cross-border payments and automatic flight booking.
Key Takeaways
- Former Anthropic researcher Jacob Coxon quit due to fears that AI development could destroy humanity by the end of the decade.
- The physical AI market is expected to grow at an annual rate of more than 30 percent over the next ten years.
- Forecasts predict over 1.3 billion AI-powered robots and autonomous systems will operate worldwide by 2035.
- The global corporate e-learning market is forecast to reach $335 billion by 2030 as companies reskill workers.
- Articulate launched Frontline, an AI platform enabling employees to create high-quality training content themselves.
- Talkdesk and Microsoft partnered to allow businesses to purchase automation tools directly via the Azure Marketplace.
- Stony Brook University received $298,000 to build Rubo, a humanoid robot for campus assistance.
- A new report warns AI automation could cause up to 800 million job losses globally by 2030.
- IBM researchers identified a 24-point gap in AI agent consistency and are developing memory-based solutions.
- Fintechs at GFF 2026 launched Agentic AI platforms and UPI Tap & Pay for payments without internet.
AI researcher quits over fears of human extinction
Jacob Coxon, a researcher at Anthropic, recently quit his job because he believes AI could destroy humanity. He stated that AI developers earnestly think super intelligent agents might kill everyone by the end of the decade. Despite these warnings, major AI companies continue to race to build better models without slowing down. Coxon argues that the current safety plans are not good enough and urges the industry to stop this dangerous competition.
Anthropic researcher warns AI race threatens lives
Former Anthropic researcher Jacob Coxon told NPR that AI development poses a danger greater than any other human activity. He warned that the current race to build super intelligent agents could threaten human existence by the end of the decade. Coxon is urging big AI companies to slow their development to prevent this potential disaster. He discusses his fears and hopes for changes in how these powerful technologies are created.
Physical AI market expected to grow by 30 percent
The next big opportunity in AI involves Physical AI, which applies artificial intelligence to the physical world. The market is expected to grow at an annual rate of more than 30 percent over the next decade. Forecasts predict that by 2035, more than 1.3 billion AI-powered robots and autonomous systems will operate worldwide. These systems will help solve labor shortages and manage complex tasks in industries like manufacturing and infrastructure.
AI infrastructure focus shifts to data movement
AI infrastructure is changing focus from just hardware to how well data moves through the entire system. While GPUs and custom chips remain important, networking and storage now play a bigger role in performance. Gartner predicts global AI spending will hit about $2.59 trillion, with nearly half dedicated to infrastructure. Experts warn that simply buying more hardware will not solve issues if data cannot flow smoothly between systems.
AI tools help companies create training content faster
The global corporate e-learning market is forecast to reach US$335 billion by 2030 as companies need to reskill workers quickly. Traditional methods of creating training content are too slow because they rely on specialist teams to translate expert knowledge. Articulate has launched Frontline, an AI-powered platform that allows employees to create high-quality learning content themselves. This tool helps subject-matter experts turn documents and videos into training kits without replacing professional learning teams.
Talkdesk and Microsoft partner for AI automation
Talkdesk and Microsoft have formed a new partnership to help businesses automate contact centers. Companies can now buy Talkdesk Customer Experience Automation tools directly through the Microsoft Azure Marketplace. This deal makes it easier to set up AI systems using money already spent on Azure. The collaboration lets firms modernize their customer service without changing their current technology setup.
Stony Brook creates Rubo robot for campus help
Stony Brook University received $298,000 to build Rubo, a new AI robot for libraries and campuses. This project is one of the first in the world to combine a humanoid robot with advanced AI conversation skills. Rubo will help students find their way, answer questions, and get support in person at high-traffic areas. The robot uses local AI models and will assist with tasks like student onboarding and multilingual support.
AI automation could cost 800 million jobs by 2030
A new report warns that artificial intelligence could cause up to 800 million job losses worldwide by 2030. This threat affects both employment and overall economic growth as machines replace human workers in many industries. Experts suggest that governments and schools must invest in education to help people learn to work alongside AI. New policies supporting green energy and sustainable farming could also create fresh jobs to replace lost ones.
IBM finds AI agents fail to stay consistent on tasks
IBM researchers discovered a 24-point gap in how often AI agents succeed when running the same task multiple times. Their study of the AppWorld benchmark showed that while average success rates look good, agents often fail to complete all five required steps. The team found that standard methods cannot fix this instability because the underlying probabilities of the AI models still vary. They developed a new system that uses memory and guidelines to improve consistency and make agents more reliable for real-world use.
Fintechs launch AI tools and new payment features at GFF 2026
Several Indian and global fintech companies launched new products at the Global Fintech Fest in Mumbai from September 8 to 11. The Department of Financial Services introduced open-source Android ATMs that can open accounts and issue cards in minutes. NPCI launched an Agentic AI platform and UPI Tap & Pay, which allows payments without internet or a phone PIN for small amounts. Other companies like PhonePe, BharatPe, and Pine Labs added AI features for cross-border payments, credit coaching, and booking flights automatically.
Sources
- ‘We really do earnestly believe AI could kill all humans’: if AI labs are so worried about AI doom, why don’t they stop?
- Are big AI companies gambling with our lives?
- When AI meets reality, the real revolution begins
- AI infrastructure shifts focus from hardware to data movement and system resilience
- AI decentralising corporate training as demand for workplace learning surges
- This Week In Cloud AI - Talkdesk and Microsoft Partner to Enhance AI in Contact Centers
- Get Ready to Meet Rubo: A Next-Generation Embodied AI Robot for Library and Campus Support
- WHAT THE TECH? AI threat
- Closing the Consistency Gap: Self-Evolving Agents That Learn to Stay on Course
- Agentic AI and other products rolled out by fintechs at GFF 2026
Comments
Please log in to post a comment.