From virtual catwalks to billion-dollar data bets, this week’s headlines show how AI is rewriting the rulebook – and why the legal fine print matters more than ever. In the UK, a Vogue ad sparks controversy for featuring an AI-generated model with barely visible disclaimers, raising transparency and diversity concerns. Meanwhile, Gulf states are racing to turn data into the ‘new oil,’ but face tough challenges around regulation, trust, and talent gaps. And in Switzerland, a fresh approach to open-source AI offers a glimmer of hope for a more ethical, transparent future.

In this edition:

📸 Strike a Pose: AI Models Walk the Legal Line

🛢️ The Gulf Bets Big on AI as it Seeks the ‘New Oil’

🇨🇭 Alp-ine Intelligence: Switzerland’s Open-Source AI Takes the Stage

Strike a Pose: AI Models Walk the Legal Line

A fully AI-generated model has appeared in the August 2025 print edition of Vogue, marking a first for the fashion magazine. The computer-generated figure featured in a paid advertisement for clothing brand Guess, created by Seraphinne Vallora, a company specialising in digital modelling. The model was not commissioned by Vogue editorial staff.

Although Guess included a note that the model was AI-generated, the disclaimer is small and easily missed. At a glance, the image appears entirely lifelike. Without clearer labelling, readers may assume they’re looking at a real person. The lack of transparency has sparked concern – particularly as UK law does not currently require advertisers to disclose when content is created using artificial intelligence.

The campaign has sparked widespread debate across the fashion industry and among digital rights advocates. Critics argue that such developments risk undermining diversity and misleading consumers. Former model and tech commentator Sinead Bovell previously warned in Vogue that AI could displace real models, disproportionately affecting those from underrepresented groups. Felicity Hayward, a body-positivity advocate, called the move “disheartening” and “a step backwards” for representation.

Health campaigners have also raised concerns. The UK-based eating disorder charity Beat warned that exposure to digitally perfected bodies could worsen body image issues, especially among younger audiences. This incident highlights growing gaps in UK regulation around AI-generated content in advertising. As AI continues to blur the line between real and synthetic media, legal questions surrounding transparency, consumer protection, and psychological harm are becoming increasingly urgent.

🔴Misleading Ads? Risks under UK Consumer and Advertising Law

If AI-generated models look real but aren’t clearly labelled, advertisers could be breaching rules on misleading content. Under the UK’s CAP Code and the Consumer Protection from Unfair Trading Regulations 2008, ads must not deceive consumers—especially by leaving out key information. In this case, a small, easy-to-miss AI disclaimer may not be enough. If a viewer assumes the model is human, the ad could be considered misleading. The ASA (Advertising Standards Authority) may step in, and brands could face reputational or legal fallout – especially where ads impact vulnerable audiences, like young people.

🟡 No AI Labelling Rules – Yet

Unlike the EU, the UK doesn’t currently require advertisers to say when content is AI-generated. The ASA recommends transparency, but there’s no legal duty to disclose. That creates a grey area: brands can technically stay silent, even when digital creations mimic reality. As AI use grows, this lack of clarity may prompt calls for reform. Parliament could eventually introduce labelling rules to ensure viewers know when something, or someone, is synthetic.

🟢 Chance to Shape AI Rules for Ads

This case exposes a wider regulatory gap: the UK has no clear laws on AI use in advertising. That gap could be addressed through future laws, such as the Digital Markets, Competition and Consumers Bill or AI-specific legislation. Introducing fair, proportionate labelling rules could strike a balance – protecting consumers while allowing innovation. The fashion industry, known for influencing body image, could be a key test case for future AI regulation.

The Gulf Bets Big on AI as it Seeks the ‘New Oil’

Oil-rich Gulf states are laying the groundwork for a new kind of energy boom – powered not by fossil fuels, but by data. Saudi Arabia and the UAE are positioning artificial intelligence (AI) as their next economic frontier, hoping to transform silicon into sovereign strength.

In Saudi Arabia, the Public Investment Fund has launched Humain, a state‑owned AI company slated to build factories powered by hundreds of thousands of Nvidia chips. Abu Dhabi is pursuing a parallel path via the MGX AI fund, and tied into projects like the US–UAE “Stargate” AI centre.

The timing serves both urgency and opportunity. As forecasts place global oil demand peaking by the decade’s close, Gulf nations aim to insulate their economies by monetising AI, with McKinsey estimating a potential $150 billion boost to its GDP.

Yet the shift brings challenges. Gulf economies have a mixed record on diversification. The host‑data‑centre model, while lucrative, risks diverting vital resources – energy and water – and may not deliver sustainable growth. Instead, officials are targeting strategic industries where AI can be embedded: manufacturing, port operations and energy infrastructure. Saudi Aramco, for instance, already uses machine learning to detect pipeline blockages.

Talent remains a major bottleneck. Though the UAE draws foreign experts with low taxes and “golden visas”, domestic education and skills gaps mean local supply falls far short of demand – particularly in data science and machine learning.

Trust is also paramount: Gulf countries must build robust regulatory frameworks and data governance protocols. Without transparency, foreign firms may hesitate to host data or collaborate on AI innovation under governments viewed as opaque or authoritarian.

🔴 Data Sovereignty & Cross-Border Transfer Risks

As Gulf states build sovereign AI models and vast data centres, concerns are growing over where data is stored and who controls it. UK and EU companies must tread carefully – transferring personal data to countries with weak privacy laws or broad state surveillance powers could breach UK GDPR, particularly Article 48, which limits data disclosures to foreign governments. Without strong legal safeguards in place, such transfers may be challenged by regulators or the courts.

🟡 Algorithmic Bias & Accountability Gaps

The rapid rollout of state-backed AI in the Gulf – especially in sectors like policing, infrastructure and logistics – raises legal red flags around discrimination and transparency. If Gulf-developed systems are deployed in UK-facing services or supply chains, they could fall foul of UK equality laws and Article 14 of the ECHR, particularly where decisions affect protected groups. Without explainability and auditability, these systems risk undermining public trust and legal compliance.

🟡  Intellectual Property & Export Control Issues

When public money funds private AI ventures, IP ownership becomes a legal grey area. UK companies collaborating with Gulf-backed projects must guard against IP disputes, trade secret misuse, and inadvertent breaches of export control laws, especially where the tech has dual-use potential. Clarity over licensing, use, and jurisdiction will be essential.

🟢 A Chance to Shape Ethical AI Globally

There is a strategic opportunity for the UK to lead in cross-border AI governance. By working with Gulf partners on shared standards for data protection, algorithmic transparency and ethical AI, the UK could strengthen both its global influence and its legal safeguards at home.

Alp-ine Intelligence: Switzerland’s Open-Source AI Takes the Stage

Switzerland is set to enter the artificial intelligence arena with the launch of its own open-source large language model (LLM), aiming to provide an alternative to major US and Chinese players. The new initiative, led by a consortium of Swiss research institutions and tech firms, intends to promote transparency, data privacy, and greater control over AI development.

The Swiss LLM will be publicly accessible and open-source, allowing developers and researchers to inspect, modify, and improve the model. This approach contrasts with many proprietary models, where the underlying code and training data remain confidential. By embracing openness, Swiss developers hope to foster innovation while addressing concerns about data security and algorithmic bias.

Switzerland’s move also reflects growing European interest in developing independent AI technologies that comply with regional data protection laws, such as the General Data Protection Regulation (GDPR). The project will be built with privacy and ethical use at its core, with strict guidelines on data sourcing and user consent.

The launch positions Switzerland as a key player in the evolving AI landscape, providing a model that balances cutting-edge technology with legal and ethical standards. For UK lawyers and tech professionals, this development signals a shift towards more transparent AI tools that respect data rights, potentially influencing future regulation and AI adoption across Europe.

As AI continues to reshape industries, initiatives like this offer a blueprint for responsible innovation in a heavily regulated environment.

🔴 Downstream Harm & Algorithmic Accountability Gaps

Once released, an open-source LLM can be fine-tuned or deployed in high-risk sectors (e.g. legal advice tools, health chatbots) by third parties. If bias, hallucination or unsafe outputs arise, pinpointing liability becomes complex. While Swiss developers may argue they’re not responsible for downstream use, this grey area is increasingly challenged by EU AI Act proposals and UK AI assurance frameworks. Regulators could still view insufficient usage restrictions or warning notices as negligent deployment.

🟡 Model Licensing & Open Source Compliance Risk

If the Swiss LLM adopts a permissive licence, commercial users may assume minimal restriction. But unclear usage terms, particularly around model fine-tuning, redistribution, or safety guardrails, could lead to IP disputes or breach of licensing conditions. In the UK, where open-source compliance isn’t yet a regulated field, this remains an overlooked legal blind spot – especially for startups rapidly building on open models.

🟢 Legal Sandbox Potential for Trustworthy AI

Switzerland’s rights-focused model aligns with emerging trends in “trustworthy AI” regulation. By embedding privacy-by-design and publishing model cards or transparency audits, the project could serve as a legal sandbox for compliant AI development. For UK firms wary of proprietary LLM risk exposure, this may present a safer, auditable alternative – particularly under future UK DRCF or EU AI Act guidance.