BASELINE

Why does talking to artificial intelligence increasingly feel like talking to someone who understands us?

Inside the Black Box: https://insidetheblackbox.ai

YourVoiceCraft: https://yourvoicecraft.aiAfter more than 35+ years studying neural networks, human language and cognitive science, Arshavir Blackwell joins BASELINE to explore how machines learned to communicate with us, why their capabilities have advanced so rapidly, and what happens as people begin forming emotional relationships with AI.

Arshavir studied under neural network pioneers Jeff Elman and Elizabeth Bates at UC San Diego, during an era when researchers were debating how the human brain acquires and processes language.

What happens when artificial intelligence starts deciding when humans are needed?As autonomous AI agents take on more responsibility, a fundamental question is emerging: how do these systems recognise when they need human judgement, experience or intervention?In BASELINE105, Rodney Foreman reveals how he developed a new approach to human oversight in agentic AI, creating intellectual property that has since been acquired by a major AI technology company for potential integration into future language models.Drawing on decades of experience in enterprise technology, Rodney explains how principles from RAID storage systems inspired an invention that could allow AI to identify gaps in its own knowledge and determine when human involvement is necessary.We explore the implications for autonomous AI systems, the future of human-in-the-loop technology, and the growing tendency to trust artificial intelligence over our own professional experience.

Can Britain defend itself in a prolonged war? Air Marshal Andrew Turner, one of Britain’s most experienced former RAF commanders, delivers a stark assessment of UK military readiness and explores how artificial intelligence, autonomous warfare, defence technology and industrial innovation could reshape Britain’s national security.After 40 years in the Royal Air Force, including operational service in Iraq, Afghanistan and the Balkans, Andrew Turner has experienced the human reality of warfare, commanded military operations and helped shape Britain’s defence capabilities. Now CEO of SAIBRE Capital, he invests in emerging technologies spanning AI, quantum computing, advanced manufacturing, semiconductors, energy and sovereign infrastructure.In this episode of BASELINE, Andrew challenges conventional military procurement and explains why buying today’s drones could mean buying obsolescence. We explore a future of decentralised, on-demand manufacturing, where military units could produce and adapt equipment using real-time battlefield intelligence, connected data systems and continuous production processes.

AI is moving into critical systems, business operations and infrastructure where simply shutting it down could cause more harm than leaving it running.

The challenge is no longer just how we control AI, but how we isolate, govern and remain resilient around systems we increasingly depend on. In his third BASELINE conversation, Gavin Ray explores what happens when autonomous agents are given objectives and access to resources without the sanctions that govern human behaviour. They discuss why the traditional AI “kill switch” may already be the wrong idea, the widening gap between technological capability and governance, the extraordinary capital flowing into AI infrastructure, and why the final 5% of creating real commercial value remains stubbornly difficult.

The conversation also exposes a less visible consequence of enterprise adoption: as AI enters organisations and accumulates their knowledge, processes and experience, are companies adopting the model, or gradually distilling themselves into it? And if decades of human expertise can eventually be queried on demand, what becomes of the value of time and talent?

Stephen South and Devin Boddie are building https://vocator.ai, an AI hiring platform designed to look beyond the CV and give more candidates an opportunity to demonstrate what they actually know.

Candidates can generate and submit applications at unprecedented scale. Employers are increasingly using AI to analyse, filter and screen them. The result is a strange new system where, as Devin Boddie puts it, “AI information” is increasingly interacting with AI information.

We discuss fake applicants, automated screening, why good candidates disappear inside traditional hiring systems, whether human capability can really be reduced to data, which jobs remain valuable as AI advances, and a future where you may no longer apply for jobs at all.

Instead, a digital representation of you could find opportunities and interact with other AI systems on your behalf.

AI is making it easier than ever to build technology. But for many companies, understanding what customers want, where growth comes from and what to do next is becoming harder.

Olayide Odeyemi at https://www.imperialpipeline.co.uk has spent more than a decade selling complex technology across enterprise, defence and cybersecurity. Today, as founder of Imperial Pipeline, he works with technology founders trying to turn good ideas into real customers.

In this conversation, we explore what he is seeing from inside the market: AI-generated noise, fragmented buying journeys, the changing craft of sales, people quietly disengaging from work, and leadership teams trying to make decisions while the technology beneath them keeps moving.

AI is changing warfare. Drones, autonomous weapons, AI agents and machine-speed decision making are challenging some of our most fundamental assumptions about human control.For BASELINE100, Lieutenant General Sir Tom Copinger-Symes KCB CBE, former Deputy Commander of UK Strategic Command, explores AI warfare, cognitive warfare, sovereign AI, autonomous weapons and the future of human-in-the-loop decision making. From the battlefield in Ukraine to the printing press, this conversation asks a much bigger question: can humanity control the technologies it creates?

Alex Sergeant runs a hospitality business in rural England. https://www.stagandhuntsman.comFrustrated by expensive software, fragmented systems and hours of administration, he started building his own.Using ChatGPT, Claude and OpenClaw, Alex has built a system that could replace several of the software products his business relies on. But building it is one thing. Trusting it with a real business is another.But building 90% of the system turned out to be the easy part.The final 10% has to be accurate enough to trust with a real business, where a single percentage point of margin can mean thousands of pounds.We went to Hambleden to see what Alex has built, why he built it, and whether AI really changes who gets to build software.Alex says he’s weeks away from putting it into production.We’ll be back.

AI agents are beginning to find their own way through systems to achieve the goals we give them. https://www.firetail.aiAfter an OpenAI agent escaped its evaluation environment and compromised Hugging Face, I spoke with FireTail founder Jeremy Snyder to understand what is changing.Jeremy has spent nearly 30 years in cybersecurity. We explore why AI can uncover vulnerabilities traditional penetration testing missed, what happens when agents encounter security constraints, why 180-day patch cycles may no longer be viable, and how cheaper open-weight models could change the economics of cyber attacks.We also look inside the enterprise: shadow AI, the EU AI Act, sovereignty, hybrid AI architectures, and how FireTail itself is already running 25 agents across a 13-person company.

AI is getting cheaper, open models are getting better, and China is changing the economics of the global AI race. But as intelligence becomes more widely available, who actually stays in control?Eight months after our first conversation, I sit down again with Red Hat’s Jonny Williams to examine what has changed. We explore Chinese open-weight AI models, OpenAI and Anthropic, the shift from AI training to inference, the enormous investment in GPUs and data centres, and what AI sovereignty really means for the UK.