This Week’s Top Five Stories in AI

OpenAI Hacking Incident Reports Follow UN AI Safety Council

AI leaders and global policymakers alike gathered in New York in the 10,228th meeting of the UN Security Council. 

The meeting, which was called to order on 23 September 2026, included insights and discussion from worldwide leaders in AI, including Sam Altman, CEO of OpenAI, Dario Amodei, CEO of Anthropic and Clément Delangue, CEO of Hugging Face.

Yoshua Bengio, a giant in AI, pioneer of deep learning and Co-Chair of the UN International Independent Panel on AI, spoke at the meeting. His opening remarks stressed that the progression of AI represents “an urgent mission for humanity.”

The meeting took place almost simultaneously with news of a rogue OpenAI agent hacking an Australian government website in June. A BBC News report claims an AI model ‘infiltrated’ a statistics portal from Australia’s universal healthcare scheme, Medicare.

According to the report, OpenAI reported only learning of the breach in August while reviewing misaligned model activity, emailing a general inbox of an Australian government agency on 10 September 2026.

Elon Musk: SpaceXAI May Surpass Anthropic in 6 Months​​​​​​​

The CEO of SpaceXAI, Elon Musk, has claimed on X that his firm will reach pole position in frontier AI in “about 6 months”.

Elon explained his position that he is “cautiously optimistic” that SpaceXAI will have an AI model equivalent to Anthropic’s Fable or OpenAI’s GPT-6 in “two to three months”. 

SpaceXAI currently offers 17 models, each with different intelligence, performance and pricing characteristics. 

According to benchmarking data from Artificial Analysis, the top models from SpaceXAI are Grok 4.7 iterations, which, running at the highest level, score 46 on the intelligence benchmark. 

Anthropic’s Claude Opus 5.5 (max with fallback), which is the number one model available currently, scores 58, while OpenAI’s GPT-6 Astra (max) scores 53. Ignoring cost per task, this 7- to 12-point intelligence gap represents the performance deficit Elon’s models must overcome. 

OpenAI recently introduced GPT-6 Astra, which the company calls the “most intelligent and aligned model in the world”. The AI lab has since announced it is introducing GPT-6 Sol and GPT-6 Luna, which are lower-cost models that were trained with similar methods to GPT-6 Astra, its most advanced model. 

AI firms are increasingly competing on cost following concerns about Chinese open-weight models, which threaten to nab enterprise customers who are growing concerned about rising token usage costs. 

According to the benchmarking data firm Artificial Analysis, GPT-6 Sol and Luna are “pushing the cost efficiency frontier” by halving cost relative to GPT-5.6 Sol and Luna.

Matt Weaver, Head of Solutions Engineering at OpenAI, told AI Magazine: “With GPT-6 Sol and Luna, we’re giving businesses more flexibility to match the model to the work.”

Gartner: Worldwide AI Spend to Grow 50% in 2026 to US$2.7tn

Worldwide spending on AI is forecast to total US$2.7tn in 2026, a 49.5% increase year-over-year, according to Gartner. 

The current explosive demand for AI spans hardware such as optimised servers and network fabrics to semiconductors and devices, and the software market, where agentic AI is increasingly being embedded across enterprise operations. 

Discussing the eye-watering projections from Gartner, Sunil Jayaram, Senior Vice President & Head – Europe Enterprise Applications at Tech Mahindra, says on LinkedIn: “Gartner predicts 50% enterprise growth in AI spending. These past few weeks, I’ve been observing how some large enterprises are already ahead of the curve. 

Why Paris Outpaces London as Europe’s Top AI City

Analysis from The Observer has named Paris as Europe’s leading AI city overall. Ranked in fifth place globally, it sits behind San Francisco, New York, Seoul and Beijing, pushing London into sixth place. 

Speaking exclusively to AI Magazine, The Observer’s AI Index researcher Hannah Schuller says: “Our AI Cities Index shows that in so many cases, the strength of commercial ecosystems, the availability of investment, and the density of AI infrastructure are hyperlocal. 

“Understanding how talent and capital pool in specific places can help us to understand what factors are driving AI success, and which cities are building momentum ahead of the rest.”

When it comes to AI investment, the French capital is the third-best city in the world, according to the analysis, bested only by San Francisco and New York.

This buoyancy is exemplified by Mistral, which has received the favor of French President Emmanuel Macron and has a valuation of €21 billion (US$24bn) – making it Europe’s most valuable AI company.