1. UK science teachers using GenAI reduce lesson planning time by 31%

A study comparing GenAI user and non-GenAI user groups found that ChatGPT-assisted teachers saved an average of 25.3 minutes a week on their lesson planning: from 81.5 minutes a week without AI tools down to 56.2 minutes a week when using them. This represents a time saving of 31 per cent per week.

2. Students using ChatGPT to write essays show 55% less brain connectivity than ‘brain-only’ peers

Students asked to write essays using large language models (LLMs) show 55 per cent less brain connectivity during the writing process compared to students using ‘brain-only’ methods, according to a study conducted by researchers from MIT, MIT Media Lab, Wellesley College and the Massachusetts College of Art and Design. The LLM group, the team wrote, experienced a ‘more automated, scaffolded cognitive mode’, demonstrating the lack of critical thinking skills or cognitive effort required to create written content using GenAI tools.

3. IEA: Data centre energy consumption to double by 2030, with AI the leading growth driver

The typical AI-focused data centre uses the energy equivalent of 100,000 households, with larger data centres currently under construction set to consume 20 times more energy, the International Energy Agency finds. The IEA reports that, in 2024, data centres accounted for roughly 1.5 per cent of global electricity consumption.

4. Stanford 2025 AI Index Report: Global AI optimism rising, but with ‘deep regional divides’

Global AI optimism is rising but remains divided across regions, according to Stanford University’s 2025 Artificial Intelligence Index, released in April. The report finds that AI products and services are seen as more beneficial than harmful in China (83 per cent), Indonesia (80 per cent) and Thailand (77 per cent), but that optimism is lower in Canada (40 per cent), the US (39 per cent) and the Netherlands (36 per cent). Opinions have grown more positive in Germany (+10 per cent) and Great Britain (+8 per cent).

5. Only 41% of surveyed councils have a defined AI policy

Only 41 per cent of councils have a specific AI policy, a survey by the UK’s Local Government Association in the United Kingdom found. The study revealed the top AI risks on councils’ agendas: cybersecurity (83 per cent), deep fakes disinformation (69 per cent) and organisational reputation and resident trust (68 per cent). The survey indicates that, despite a general awareness of risks, there is a significant lag on governance and policy to protect against them.

6. Wage premium for AI skills is growing, PwC finds

Job candidates with AI skills see a 56 per cent wage premium compared to those without them, according to PwC’s 2025 Global AI Jobs Barometer. This is a 31 per cent increase from last year. The study also found that the skills needed for ‘AI-exposed’ jobs are changing 66 per cent faster than for other jobs.

7. More people around the world use AI than trust it: KPMG

A global study by KPMG surveyed 48,000 people across 47 countries and found that 66 per cent use AI regularly and 58 per cent use AI at work on a regular basis. Eight-three per cent of respondents believe that using AI will provide a wide range of benefits, although only 46 per cent say they trust the technology. Sixty-six per cent also say that they rely on AI results without checking for accuracy, and 56 per cent acknowledge making mistakes in their work because of AI. A majority – 70 per cent – also see the need for national and international regulations on AI.

8. HubSpot: 55% of marketers rely on AI for writing in 2025

More than half (55 per cent) of marketing professionals who use AI rely on AI for writing, according to a HubSpot survey of people across North America, Europe, Asia and Australia. The survey also found that 43 per cent say that the biggest challenge in using GenAI is that it produces inaccurate information, 34 per cent say that outputs are sometimes biased and 31 per cent say that GenAI outputs aren’t always relevant to their goals. When they look for education on how to use AI, 56 per cent of respondents report turning to video platforms such as YouTube and TikTok, while 48 per cent say they rely on resources from their employers.

9. 82% of US business leaders believe AI will transform their industries in next 2 years

KPMG’s AI Quarterly Pulse Survey: Q2 2025 finds that 82 per cent of US-based business leaders believe that AI will transform their industry’s competitive landscape within the next 24 months. It also finds that most organisations have moved beyond experimenting with AI agents, with 33 per cent having already deployed some. Nearly half – 46 per cent – of respondents say they are focusing their AI agent strategies on improving efficiency and growing revenue.

10. Profusion of AI-ready datasets fuels ‘explosion’ of formulaic and inappropriate research studies

The growing availability of AI-ready datasets has led to an ‘explosion’ of formulaic scientific research articles, inappropriate study designs and false discoveries, a team of UK-based researchers from the University of Surrey and Aberystwyth University has found. The team found that, while just four National Health and Nutrition Examination Survey (NHANES) association-based studies were published each year between 2014 and 2021, those numbers rose to 33 in 2022, 82 in 2023, and 190 in 2024. The team recommends that researchers follow a set of best practices to address these concerns and ‘mitigate the risks of paper mills using AI-assisted workflows to introduce low-quality manuscripts to the scientific literature’.