Commercially Connected Shorts - 5 February 2025

This week we look at:

• EU: what to expect in the drive for competitiveness – the roadmap
• AI: what does the first global handbook on AI safety say?
• Public sector: Blueprint for digital government – the UK’s plans and impact for businesses
• Tech/Privacy: UK’s Online Safety Act: what service providers need to know

The EU’s roadmap to drive competitiveness

On 29 January 2025, the European Commission announced its three core areas for action focussing on innovation, decarbonisation and security.  With the aim of boosting the EU’s competitiveness (through the aptly named “Competitiveness Compass” roadmap) we can expect to see initiatives which focus on the following areas:

• Boosting innovation - “AI Gigafactories” and action plans for the increased adoption of AI in the following sectors: advanced materials, quantum, robotics, biotech and space technologies
• Access to clean and affordable energy through the Clean Industrial Deal (the EUs approach to decarbonisation, boosting manufacturing, clean tech and circular business models). Support for energy intensive sectors of steels, metals and chemicals to transition through an Industrial Decarbonisation Accelerator Act.  Addressing the challenge of high and volatile energy prices with an Affordable Energy Action Plan to bring down energy prices and costs
• Clean Trade and Investment Partnerships to help “secure supply of raw materials, clean energy, sustainable transport fuels, and clean tech from across the world” with the review of the Public Procurement rules expected to introduce a European preference for critical sectors and technologies. The aim here is to reduce excessive dependencies and increase security

These three core initiatives will be complemented by a change in approach through five “enablers” which focus on:

• Simplification – reducing regulatory and administrative burdens. We have already heard of plans to simplify the EU sustainability reporting regimes through the Omnibus proposal
• Removing intra EU barriers through a Horizontal Single Market Strategy – modernising the EU governance framework
• Turning savings into investments through a European Savings and Investments Union
• Promoting skills and quality jobs to match skills with labour market demands
• Improving policy coordination at EU and national levels with a Competitiveness Coordination Tool to identify shared objectives and investments and a Competitiveness Fund to provide financial support for this

The Compass provides us with an indication of the laws and policies which the Commission will present to the EU Parliament and Council over the coming years.  Expect:

• new laws to support SMEs and investment and focus on AI and tech
• further support to become green in manufacturing and focus on creating secure supply chains to boost economic success

Businesses operating with or in the EU should review the detail of the proposals to gain an understanding of how they can align their own strategy to benefit from any future investment and remain compliant with law / policy.

What does the first global handbook on AI safety say?

On 29 January 2025, the UK government published the First Independent International AI Safety Report.  This report will be presented at the French AI Safety Summit later this month and sets out “the first comprehensive, shared scientific understanding of advanced AI systems and their risks” having had input from 100 world-leading AI experts. 

With a scientific evidence-based focus, the aim is to present the risks of general purpose AI (i.e. AI that performs a wide range of tasks ‘gpAI’) and suggest methods for mitigation – these in turn can enhance understanding and support policy making.  So what are the key messages which might drive conversation and action at the forthcoming summit?

• the report is a moment in time capture as already advancements in gpAI continue to enhance capability (and in turn risk spotting and mitigation which help highlight the safety approach)
• company awareness and investment in gpAI technology is increasing – as more gpAI is deployed the need to manage the risks and share good practice increases so that safety is paramount in managing the capabilities of this rapidly developing tech
• gpAI models with more computing power may produce better outputs but this is still costly
• mitigations for the harms of gpAI (such as scams, abusive or intimate images and bias) are in place but there is no one perfect solution meaning gaps or more subtle forms of bias need addressing
• risks associated with gpAI include: malicious use – harms to individuals and organisations (think fake content, manipulation of opinion and cyber-attacks); malfunctions – the reliability, managing bias and loss of control challenge; systemic risks – the impact of wider gpAI deployment on the labour market, privacy and the environment
• mitigating the risks can be a challenge when developers don’t fully understand the capabilities of the gpAI, the same system can be deployed in a multitude of ways and there are information gaps between government, companies and researchers
• there is a balance for policymakers in assessing the benefits and risks of advancements on limited real-life evidence – early warning systems and shared risk management can help address this
• focus on the following questions is required: how we measure the advancement of gpAI collectively? What are sensible risk thresholds to trigger mitigations?  Greater understanding of how the gpAI system works to produce the output and how it is designed.  How the policymakers best gain access to the information on gpAI that is relevant to public safety?

These are familiar themes and the questions posed provide insight into the possible discussion focus of the next AI Safety Summit.  From a practical business perspective the report highlights the following issues:

• how can the deployment of genAI help enhance your business?
• the importance of developing a strategic approach to gpAI – what is and is not permissible within your company?
• assess the risks, talk to suppliers, understand the tech and what is being said in any contract terms in respect of gpAI
• research and share good practices for mitigating any identifiable risks and implement these
• this technology doesn’t stand still – make it a regular agenda item at board meetings!

Blueprint for digital government – the UK’s plans

As part of the UK Government’s drive to turn the UK into an AI superpower, it has launched the “Blueprint for Digital Government”. The vision is for “modern digital government” in the UK to:

• drive growth, productivity and efficiency
• create smarter public sector organisations
• provide “easier lives” for citizens

The plan envisages a key role for digital and tech businesses in this vision and identifies six key priorities including joining up public sector services, harnessing the power of AI for the public good, reforming funding of digital and technology and maximising the value and potential of public procurement of digital and technology.  It also starts to flesh out the UK Government’s AI Opportunities Action Plan and is intended to show that the UK Government is serious about AI transforming the delivery of its services - while acknowledging that government cannot do this alone and that ambition needs to be tempered by responsibility as the increased use of technology brings risk from technological threats.

Key takeaways for businesses include:

• The new UK Government Digital Service will attempt to unify the fragmented current digital picture across Government, which should make it easier for suppliers and businesses to access a single contact point of digital policy and delivery in the UK Government
• An emphasis on partnership working with the technology sector shows the challenges that the Government faces to modernise its digital delivery in an era of limited resources. There is an opportunity for businesses to make their voices heard as the UK Government will want to hear from them about their digital needs and understand their capabilities
• Funding models and procurement models for digital and technology will change, in tandem with the new procurement legislation. Suppliers to the UK Government will need to be ready to articulate their views on the efficacy of new models and to move quickly to react

For more information on the key priorities and next steps (including a new UK Government Digital & AI Roadmap in the summer) see Update on UK Government's Blueprint for Digital Government | Eversheds Sutherland.

With thanks to Philip James, Christopher Murray and Nicolette Saunders

UK’s Online Safety Act: what service providers need to know

From 17 March 2025, the Online Safety Act 2023 (“OSA”) will require that online service providers within its scope take measures to protect their users from illegal content.  Additionally, for services that are likely to be accessed by children, any content that is harmful to children. Whilst it may seem that search engines, dating, gaming and social media platforms are the primary targets, the OSA has wider reach and application than commonly perceived. Numerous online service providers, including websites or applications available over the internet that host and index user-generated content, must pay close attention to this legislation and its approaching deadlines.

The legislation establishes varying levels of requirements dependent on the nature of the service provided. It employs different categories and scales to make the obligations vary proportionate to the associated risks. Larger, higher-risk platforms are subject to more rigorous obligations compared to smaller service providers. Providers need to stay vigilant and proactive in implementing safety measures to protect their users from illegal and harmful content. Understanding what Ofcom now expects from each type of online service provider is integral for ensuring compliance with the legislation and mitigating the risk of penalties. 

Some of the immediate key considerations and actions for organisations embarking on compliance projects for the OSA in 2025 include:

• Determine if the OSA applies to your services
• Understand the key milestones and compliance deadlines
• Undertake regular risk assessments
• Note and follow the Ofcom Codes of Practice
• Design and develop systems to include safety measures
• Be transparent about the types of content allowed on your platform and explain your approach in your terms of service (or publicly available statement). Also provide users with tools to control the content they see
• Establish clear and accessible ways for users to report harmful content or issues they encounter
• Balance safety with freedom of expression in terms of user rights
• Maintain documentation of your actions and safety measures adopted and be prepared to report on compliance to Ofcom

For more information on the scope of the OSA, what illegal content means, the penalties for non-compliance, key dates and next steps see our detailed briefing: UK’s Online Safety Act: What Service Providers Need to Know | Eversheds Sutherland.

With thanks to Paula Barrett, Philip James, Jonathan Palmer, Anna Allen and Heather Farrell-Boote