Data & AI Report – AI Making Waves Again

Welcome back! 

Last month’s Data & AI Report focused heavily on AI advancements – this month is no different! From open-source breakthroughs to policy debates and autonomous innovations, here are March’ top AI stories. 

🥇 DeepSeek Dominates Again 🥇

For the third month in a row, DeepSeek is making headlines! The latest model, DeepSeek V3-0324, now leads as the top-performing non-reasoning AI, surpassing Google’s Gemini 2.0 Pro & Meta’s Llama 3.3. 

While still catching up to reasoning models, its instant response speed makes it ideal for real-time applications like chatbots. This is a major win for open-source AI, proving its growing competitiveness, and a notable success for DeepSeek despite ongoing controversies. 

🫡 OpenAI & Google Push for U.S. AI Leadership 🫡

As China advances in AI, OpenAI & Google are urging the U.S government to act swiftly to maintain its global AI leadership. Their joint recommendations include: 

✅ Stronger export controls 

✅ Increased AI investment 

✅ Infrastructure development 

✅ Government AI adoption 

Both companies stress that policy decisions made now, will shape the future of AI dominance, impacting innovation, security & economic growth. 

👋 Meet Carl: AI’s First Automated Research Scientist 👋

The Autoscience Institute has introduced Carl, an AI system capable of independently producing academic research that passes rigorous peer review. Carl’s papers were accepted at the ICLR conference, marking a breakthrough in AI-driven scientific research. 

How does Carl work, you might ask? 

• Generating ideas: Generates hypotheses by analysing academic literature 

• Experimentation: Runs code, tests ideas, and visualises data 

• Presentation: Writes polished research papers with clear conclusions 

There’s no doubt that as AI developments unfold, we’ll be seeing more Carl like creations. The academic community is going to need to adapt to this as it develops to ensure integrity, safeguarding and relevance. 

🚗 Japan’s Autonomous Vehicle Push 🚗

Nissan has launched driverless cars in Yokohama, using a system of 14 cameras, nine radards and six LiDAR sensors. Whilst Japan have been lagging behind America & China in this area so far, they are catching up now. 

By 2029, Nissan aims to have 20 fully autonomous vehicles on the road – without human intervention. With Japan’s reputation for precision engineering, expectations are high, and failure is not an option. 

Conclusion 

Another month of AI evolution of breakneck speed! Check back for next month’s Data & AI report to see how things are progressing.