Investigating the AI Industry – Online (PM)

The rise of a global AI industry has accelerated rapidly over the past decade. Research into AI systems is dominated by some of the largest and most powerful corporations on the planet and governments and companies alike are committing billions of pounds of investment towards creating the infrastructure for AI products. Meanwhile, concerns that AI is fuelling disinformation, exacerbating inequality, feeding mental health crises and undermining efforts to address the environmental crisis and climate breakdown continue to grow. Financial institutions have warned that the speculative bubble around the valuation of AI companies may well burst with catastrophic results for households, businesses and governments.

This course equips investigative journalists with foundational knowledge and skills to get behind the hype and dig out the stories around an industry with an already significant impact on society and our lives – impacts that are set to increase exponentially. The sessions provide a strong grounding in fundamental concepts and an overview of significant trends in AI research which have shaped today’s industry and then explore the investigative angles from which to approach the industry, from the societal injustice caused by outsourcing decision-making, to the exploitation and human rights issues within a vast and poorly regulated workforce, and the environmental harms from energy use, water scarcity and mineral extraction.

The aim is to encourage informed journalism which can challenge false claims from corporate and political lobbyists and expose those profiting from practices which harm the environment and violate human rights.

Technical Requirements

This course will need you to have the following software/apps/tools on your computer:

Zoom app
Camera and audio

This course will be hosted on Zoom. To find out more about how we use Zoom, please check out our Zoom InfoSec page.

Course Structure

Exercises and additional resources will be provided to supplement the training between sessions.

Important

Our training is not recorded: if you miss a session, it is lost – you cannot watch a recording of it, nor will you be allowed to attend that session at a later date.

19 October 2026 – Technical foundations and the origins of the AI industry

14:00–16:00

The session provides an overview of the commercialisation of Machine Learning systems including the creation of large-scale training datasets and standard benchmarks, innovations in design such as Convolution Neural Networks and Pre-trained Transformers. It stresses the interaction between technical advancement and the rise of companies combining access to 'hyperscale' digital infrastructure across the key building blocks of the AI industry: compute, data and models. Participants will learn key concepts in contemporary AI systems and an understanding of how the subjective choices made by system designers have real-world consequences.

20 October 2026 – Mapping the global corporate landscape of AI

14:00–16:00

This session explores the corporate landscape of the global AI industry and how it has developed through increasing government investment and the emergence of an 'AI arms race'. It covers how companies, research labs, thinktanks and policy-makers interact and takes a deeper dive into the structure of the industry, examining interdependencies between companies across the 'stack' of technologies used to produce AI models. Participants will then learn about the history of stock market and infrastructure investment bubbles and explore comparisons between the dot.com boom of the early 2000s and the feeding frenzy around AI companies in the global economy today.

21 October 2026 – Understanding AI models and their digital supply chains

14:00–16:00

The AI industry today is driven mainly by the race to produce and monetize very large-scale AI 'foundation' models. We'll look at methods for producing the models which lie behind ChatGPT and Gemini's interfaces. Where do they get their data from? Who works on data and model production and under what kind of conditions? We will explore the world of data partnerships and brokerage and investigate the role of stolen and non-consensual content as the industry's 'raw material'. The session will also cover the development of labour supply chains for different model architectures, from data labellers to evaluators and look at sources for understanding the impact of the model production process on workers.

22 October 2026 – The rise of the AI factory

14:00–16:00

The vast energy and water demands of specialist AI data centres now regularly feature in the media spotlight, but what else lies behind the rise of the AI factory? This session provides an overview of the process from the expanding mining operations in critical minerals and rare earths, to the race to secure specialist chips, the rewiring of the energy grid, expanding water use and the impact on communities. Building on knowledge from previous sessions we will analyse AI production as a form of heavy industry by looking at the 'weight' of the cloud infrastructure required to make AI models including its subsea cable networks, fossil fuel and nuclear power sources, e-waste and pollution footprint and examine case studies exploring the impact on communities at every stage in the production process.

23 October 2026 – Investigating the impact of AI models

14:00–16:00

Our final session returns to the question of bias and discrimination with a focus on the use of AI models. We will recap on how selection of training data, labelling decisions and choice of model architecture can affect AI systems' predictions, situating these issues in a longer history of investigations into data and statistical bias. Participants will explore use cases where the adoption of AI models is likely to exacerbate existing sources of inequality or create new ones such as media and advertising, scoring systems for resource allocation in public services and predictive policing. We will discuss whether it is possible to fix discrimination in AI-based systems and whether AI tools can play a positive role in empowering investigative journalism.

Trainer Biography:

Anne Alexander

Anne Alexander has worked as a researcher and trainer specialising in critical approaches to understanding AI for the last 15 years at the University of Cambridge. She recently led a two-year research collaboration with the Pulitzer Center and Watershed Investigations developing new methods for journalism using AI in remote-sensing investigations and is a co-founder and regular teacher in the Cambridge Data Schools, an intensive programme in data-driven investigations in the public interest for journalists and researchers.

Booking Form

19 October 2026 14.00–16.00 Timezone: BST (UK Time)
20 October 2026 14.00–16.00 Timezone: BST (UK Time)
21 October 2026 14.00–16.00 Timezone: BST (UK Time)
22 October 2026 14.00–16.00 Timezone: BST (UK Time)
23 October 2026 14.00–16.00 Timezone: BST (UK Time)

Timezone: BST (UK Time)

Location: Zoom meeting

Goldsmiths students (full time)*

£121

Students (full time)*

£157

Freelancers**

£314

Small Media/Education/NonProfit Organisations (<10 staff)

£423

Large Media/Education/NonProfit Organisations (10+ staff)

£568

Other Organisations

£1017

In line with our non-profit mission, our pricing operates on a sliding scale, ensuring large organisations pay more to subsidise places for smaller newsrooms, freelancers and students.

*Student places for this course are capped, due to limited capacity. Anyone registering as a student will be asked for a photo/scan of their student ID ahead of the course.

**Employed individuals who cannot have their employers pay for the course are entitled to the freelancer rate. Note that we are a small charity and rely on your honesty so please do not register as a freelancer if your employer is reimbursing you for the course.

We have a strict policy of No Refund and No Transfer of bookings.