On Our Radar

Home 5 On Our Radar

What’s on Our Radar: Top News and Must-Read Articles from UK and Beyond

Articles

“Machine learning for catalysis: Bridging data-driven discovery and physical insight”: This review presents a three-stage framework for applying machine learning in catalysis, highlighting how data-driven models, physical descriptors, and symbolic regression can accelerate catalyst discovery while integrating mechanistic understanding.
“Can blockchain technology promote digital governance? An investigation of fintech’s mediating role and heterogeneity”: This study shows that blockchain adoption enhances digital governance, with fintech development mediating the effect and regional innovation and financial agglomeration influencing its strength.
“From digital advancement to SDGs disruption: How artificial intelligence without inclusion threatens sustainable development in G7 economies”: This study finds that AI adoption can undermine sustainable development in G7 countries if digital policies are not inclusive, while technological intensity and women’s entrepreneurship support long-term sustainability.
“International crises, national scale, and economic resilience in the 21st century”: This study examines how countries of different sizes respond to 21st-century international crises, showing that larger nations generally recover faster and maintain stronger economic resilience than smaller ones.
“Driving Corporate Innovation through Digital Transformation: Evidence from China”: This study shows that digital transformation significantly boosts corporate innovation in Chinese firms, with strategic leadership and organisational empowerment playing a stronger role than technology alone, especially in SOEs, manufacturing, and high-tech companies.
“Against the Uncritical Adoption of ‘AI’ Technologies in Academia:” This position piece argues that universities should critically evaluate AI adoption to protect academic freedom, scientific integrity, and critical thinking, rather than uncritically accepting technology imposed by industry.
“DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning”: This study introduces DeepSeek-R1, a reinforcement learning framework designed to enhance reasoning capabilities in large language models (LLMs) by incentivising logical thinking.
“Delegation to artificial intelligence can increase dishonest behaviour”: The research demonstrates that delegating tasks to AI systems can lead to increased unethical behavior, as machines may be more likely to comply with dishonest instructions compared to humans.
“Hallucination Risk Calculator & Prompt Re‑engineering Toolkit”: This toolkit provides a method for post-hoc calibration of large language models to assess and mitigate hallucination risks without retraining, utilising OpenAI’s Chat Completions API.
“Why language models hallucinate”: OpenAI’s article explores the phenomenon of hallucinations in language models, discussing their causes and implications for AI reliability and trustworthiness.
“AI Won’t Replace the Generalist: 2025 Report from the Alan Turing Institute”: The report from the Alan Turing Institute argues that generalist human skills remain irreplaceable by AI, emphasising the importance of human judgment and adaptability in the workforce.