Artificial Insights

January 2025 saw explosive growth in several AI sectors. Marketing saw a boom in AI-powered TikTok advertising.  Software development benefited from AI code assistants, while creative industries embraced AI voice generators and presentation makers.

Cutting-edge advancements in artificial intelligence (AI) are transforming numerous sectors. Significant progress is evident in customer service, search technology, and scientific research, with AI improving efficiency and accuracy.

A new machine learning algorithm accurately identifies the top five flavour notes in different whiskies, matching the assessments of expert human tasters. 

Microsoft's AI for Good Lab has developed SPARROW, a solar-powered, AI-driven device designed to monitor biodiversity in remote locations.

Researchers at Imperial College London are developing an AI-powered t-shirt to improve the diagnosis of inherited heart conditions. Funded by the British Heart Foundation, the shirt uses embedded sensors and AI to monitor heart rhythms continuously, potentially detecting abnormalities missed by traditional hospital scans. 

A UK NHS trust is utilising an AI-powered app, Lenus, to manage chronic obstructive pulmonary disease (COPD) in patients. 

Research from Imperial College London demonstrates how an AI model, trained on millions of ECGs, can accurately predict heart disease, heart attacks, and even mortality risk.

Key areas highlighted include healthcare, utilising AI for diagnostics and robotic surgery; robotics, showcasing innovative designs for various applications; military and defence, noting AI integration into weaponry and strategy; and environmental monitoring, employing AI for climate analysis and ecological observation. 

The NHS in England is employing artificial intelligence to identify and support individuals who frequently use A&E services. Over 360,000 patients visit A&E more than five times annually, often due to underlying social and economic factors. 

Google has unveiled Gemini 2.0 Flash, a significantly enhanced version of its AI assistant.  Key improvements include enhanced multimodality—processing and generating various media types—and the ability to utilise external tools like Google Search. 

Five ways artificial intelligence (AI) will impact workplaces in 2025.

Researchers at the John Innes Centre and the University of Exeter have developed PlantRNA-FM, a groundbreaking AI model that analyses the genetic "language" of plants.

Announcing OpenAI's release of its new language model, ChatGPT o1, touted as the "world's smartest". 

Several key trends in artificial intelligence, including the rise of multimodal and agential AI systems, the increasing prevalence of shadow AI within workplaces, and significant advancements in AI-powered robotics and healthcare.

Google recently launched Veo, its first AI video generator, and Imagen 3, an advanced image generator, both accessible through its Vertex AI platform. 

iOS 18.2, an upcoming Apple software update, is expected to launch around 9th December and will feature significant AI enhancements.

AI-powered tools are improving the detection of bone fractures and early disease signs, aiding ambulance needs assessment, and assisting clinicians via chatbots.

France's plan to introduce AI-powered speed cameras in 2025, capable of detecting various driving offences beyond speeding. 

A recent report reveals a disparity between the hype surrounding AI in the workplace and its actual adoption. While 60% of companies claim AI tool usage, a significant portion of employees (34%) use such tools less than monthly.

The UK government is establishing the Laboratory for AI Security Research (LASR) to combat the growing threat of AI-enabled cyberattacks. 

New research from Ziff Davis reveals that leading AI companies, including OpenAI, Google, and Meta, heavily rely on high-quality content from premium publishers to train their large language models (LLMs), despite downplaying this reliance. 

Apple is reportedly developing a next-generation Siri, codenamed "LLM Siri," based on large language model (LLM) technology. This upgraded Siri, aiming for a spring 2026 release, will allow for more human-like conversations and integration with third-party apps, similar to ChatGPT.

A new AI system, Squirrel Agent, is being trialled in the UK to help conserve red squirrels, whose numbers are dwindling due to competition from grey squirrels and a lethal virus. The AI identifies red and grey squirrels with 97% accuracy, automatically controlling access to feeders dispensing food or contraceptives.

Virgin Media O2 in the UK has deployed an AI-powered "granny" named Daisy to combat phone scams. Daisy engages scammers in lengthy conversations, wasting their time and frustrating them with irrelevant details. 

In this podcast we discuss recent key developments in artificial intelligence.  Significant advancements are highlighted across various sectors, including the release of new AI agents and assistants (Operator and Gemini), substantial progress in AI-powered robotics and healthcare applications, and the emergence of new AI platforms.

The Stanford Institute for Human-Centered Artificial Intelligence (HAI) has released a Global AI Vibrancy Tool, ranking countries based on their AI ecosystems. The tool assesses eight pillars, including research, economy, and policy, using 42 indicators. 

Sir David Attenborough has expressed his deep concern about the use of AI to create a clone of his voice. Attenborough objects to his voice being used to express opinions that he does not hold. 

This episode discusses the potential extinction of hyperlinks due to the rise of AI. It traces the history of hyperlinks, from Vannevar Bush's "memex" concept to the legal battles surrounding web crawling and Section 230.

Researchers at Washington State University have developed an AI deep learning model that can rapidly and accurately identify signs of disease in tissue images, outperforming human pathologists. 

A report from Canalys reveals that a significant portion of PCs shipped in the third quarter of 2024 are equipped with AI capabilities. These AI-optimised PCs feature dedicated chips, known as neural processing units (NPUs), designed to accelerate AI tasks directly on the device.

Generative AI is becoming more widely adopted across industries, leading to improvements in customer service, content creation, and operational efficiency.

Apple has introduced new AI-powered features to its video editing software, Final Cut Pro 11. 

The NHS in London is trialling a generative AI assistant called Tortus to improve patient care.  Tortus uses ambient voice technology, allowing doctors to listen to patients and focus on them rather than typing notes.

We discuss the development of new rules within the Robots Exclusion Protocol aimed at preventing AI training bots from using public web content without publishers' permission. 

A recent report suggests that OpenAI's next-generation language model, Orion, is facing significant challenges. Training data scarcity and the increasing cost of training are hindering progress, leading to slower development than anticipated. 

Microsoft is introducing new AI-powered features to its Paint and Notepad applications for Windows Insiders. The new Generative Fill feature in Paint allows users to add or edit content using text prompts, while Generative Erase lets users remove unwanted objects from images.

The latest trends and applications of artificial intelligence (AI) across various sectors.  We highlight key advancements in generative AI, robotics and automation, scientific discovery, healthcare, and business. 

A recent study from the University of Kansas Life Span Institute has found that parents are increasingly turning to AI, specifically ChatGPT, for health advice regarding their children, even rating the AI chatbot as more trustworthy and accurate than human doctors. 

Data quality is crucial for the success of artificial intelligence (AI) systems.

The Institute of Job Aggregation (IJA) has released a report highlighting the growing concerns surrounding the use of artificial intelligence (AI) in recruitment. The report reveals that nearly half of job seekers have witnessed an increase in fake job advertisements since AI tools have become more prevalent. 

The podcast explores the potential for manipulation of large language models (LLMs) used by AI chatbots. It discusses how these models can be influenced by “generative engine optimisation” (GEO), a practice similar to search engine optimisation (SEO) that aims to increase the visibility of certain content in chatbot outputs.

Amazon plans to expand Alexa’s capabilities by deploying it as an AI agent, allowing it to perform tasks independently. This development follows Amazon’s announcement that Alexa would be equipped with generative AI technology. 

The latest developments in artificial intelligence (AI), highlighting the prevalence of generative AI and its applications in various fields.

OpenAI is reportedly developing its next major artificial intelligence model, codenamed Orion, with plans to release it by December. Unlike previous models like GPT-4, Orion will initially be made available to select companies for development of their own products and features

Apple has released the first set of its new AI features for iPhones, iPads, and Macs, called Apple Intelligence, which aims to provide users with a more intuitive and personalised experience.

The US Treasury is increasingly using artificial intelligence (AI) to fight fraud, which has helped them recover billions of dollars. The technology has been instrumental in identifying and preventing improper payments, with a significant increase in fraud prevention compared to the previous year. 

This podcast explores how AI-powered software development is transforming medical practices in the UK.

Meta, the owner of Facebook and Instagram, is introducing facial recognition technology to combat celebrity impersonation scams. These scams often promote fraudulent investment schemes and cryptocurrencies using deepfake technology to create realistic likenesses of celebrities. 

In this episode we discuss the trending topics in AI this week such as AI search and language models with companies developing more natural, conversational search capabilities.  Multimodal AI, which combines different data formats, is gaining traction, allowing for more human-like interactions.