Artificial Intelligence (AI) is rapidly transforming the business landscape worldwide. For senior business leaders, the imperative is clear: understand AI's potential, navigate its risks and implement it responsibly to drive value and maintain compliance.

This article provides an overview to help you make informed decisions and position your organisation for success in the age of AI.

What is AI?

AI is best described as the use of digital technology to create systems capable of performing tasks that typically require human intelligence, such as visual perception, speech recognition, decision-making and language translation.

According to the Organisation for Economic Co-operation and Development (OECD), an AI system is a machine-based system that, for explicit or implicit objectives, infers from input data how to generate outputs (predictions, content, recommendations, or decisions) that can influence physical or virtual environments. These systems vary in their autonomy (the degree to which they operate without human intervention) and adaptiveness (their ability to evolve after deployment).

How does it work?

At one level, AI is a very clever word-predicter and pattern-spotter. Based on thousands of documents, webpages, articles, books, blogs etc that the AI has ingested, it creates mathematical models of the relationship between words so it can anticipate what word comes next in a particular context. How it does that varies between different types of models and learning techniques, but it reveals something critical – AI does not 'understand', it does not have 'logic', it does not appreciate 'time' and it certainly has no 'morals'. So, whilst it is very sensitive to choice of language, it can learn itself and it can produce outputs independently, it is certainly not infallible.

Opportunities: how AI can transform your business

AI offers a spectrum of opportunities for organisations, from enhancing decision-making to automating operations.

The general benefits it delivers are:

  • Efficiency: automating routine tasks to free up human resource.
  • Accuracy: in performing routine tasks, or in monitoring live situations where a human operator could be distracted or tired.
  • Insights from data: ability to process more data than a human could to give insights.

The use cases to realise these benefits are hugely varied depending on the nature of a business. You could think about it split between generic tasks that any corporate undertakes and then operation specific uses.

General corporate uses

  • Speech-to-text: Automate meeting minutes and identify action points.
  • Document summarisation: Quickly extract key insights and options from large volumes of information.
  • Information sifting: Automated review of standard documents (e.g. CVs) to check for objective requirements.
  • First draft creation: Creating a first draft of a business case, presentation, customer communications.

Operational applications

  • Predictive maintenance: Anticipate machinery issues before they occur.
  • Workflow optimisation: Reduce idle time and streamline processes.
  • Energy management: Monitor and optimise energy usage.
  • Quality control: Enhance product and service standards.
  • Safety monitoring: Real-time oversight in physical environments.
  • Demand forecasting: Use historical data to predict future needs.

The most important thing to recognise is that identifying the right use cases is fundamental. Like the internet before it, AI is a very general tool which could be used in many different ways. There are many solutions being developed, but not every solution can resolve a problem in every business. Each organisation must assess where AI technologies can deliver the greatest benefit, tailored to its sector, its weaknesses and its strategic objectives.

Risks: what to watch out for

AI introduces risks that differ significantly from traditional software. Key concerns include:

  • Intellectual property (IP) ownership: Who owns the outputs generated by AI? Did the AI developer have all the rights and licences to use the data to train its model?
  • Unexpected results and hallucinations: AI may produce inaccurate or misleading outputs. It can make facts up, and then double down when asked if a fact is correct. It does not know to check for the most recently-dated publication or facts.
  • Black box effect: Lack of transparency in how decisions are made. AI models are getting better at offering sources so users can verify the outputs and at showing their working.
  • Confidentiality: Risks of sensitive data exposure. Unless structured to operate within an organisation's IT estate, models ingest all inputs and, if prompted correctly, could provide an input as an output to another user.
  • Discrimination and bias: AI trained on biased data can perpetuate unfair outcomes. AI 'learns' from the data it is given. It does not know if the underlying data is correct or fair.
  • Deepfakes: Manipulated media that can damage reputation. Whilst experts carry out audits on code to assess whether it has been produced by AI, these efforts still lag behind those looking to create fakes to commit fraud, spread misinformation or create fear.
  • Cyber risk: Increased vulnerability to attacks. AI is yet another digital service that can increase an organisation's attack surface. There are specific AI attacks which inject malicious code designed to cause a model to operate in unexpected way.
  • Energy consumption: AI systems are very energy-intensive compared to cloud computing. This presents a challenge for organisations who are also trying to decarbonise their supply chains.

Mitigating these risks requires a tailored approach for each use case. For example, marketing teams may focus on IP concerns, HR on discrimination and customer-facing functions on hallucinations and reliability.

The rules: navigating UK and European AI regulation

UK regulation

The UK government has adopted a pro-innovation, principles-based approach, leveraging existing regulators and legislation. While there is no AI-specific law yet, all current UK laws apply, including:

  • Data Protection (GDPR)
  • Equality Act
  • Consumer protection
  • Copyright laws

The government's five guiding principles are safety, security and robustness, transparency and explainability, fairness, accountability and governance, and contestability and redress. These aim to build public trust and ensure proportionate, context-specific regulation. There are indications that the King's Speech in 2026 will announce AI legislation designed to prohibit the most harmful uses of AI, protect creators of IP but also foster innovation – a difficult balance to strike.

European regulation

Europe is leading with the EU AI Act, which complements existing legislation such as GDPR, the Digital Services Act, and the Digital Markets Act. The AI Act uses a risk-based model, categorising AI applications by potential harm and imposing strict legal obligations on developers and users. The AI Liability Directive and a proposed EU AI Board are also in development.

Global perspective

Other jurisdictions, including the US, China and Canada, are developing their own frameworks, often focusing on specific applications or ethical considerations. For multinational organisations, cross-jurisdictional compliance is essential.

Responsible AI: what are others doing?

Leading organisations are adopting "responsible AI" principles, including:

  • Transparency and explainability: Disclose AI use and provide information on how it works.
  • Fairness and non-discrimination: Address bias and ensure equality.
  • Robustness, safety and security: Monitor for unintended or malicious use.
  • Accountability and governance: Establish clear oversight and the ability to override AI decisions.
  • Active risk management: Demonstrate ongoing consideration of risks and governance.

Many companies set up steering groups and policies to ensure their AI usage stands up to scrutiny, both legally and ethically. AI is challenging for governance because it is multi-disciplinary – being a sophisticated digital tool, with specific cyber, data and legal risks, which could be used in any operational part of a business where it is crucial to have stakeholders from those functional areas involved to understand the problem to be solved and define what success looks like.

It is also important to decide what approach to take to AI used by employees. Should employees be prohibited from using AI except for tools deployed centrally which are permitted? Or should employees need to get permission to subscribe to or use AI tools per use case? Or could employees be permitted to use any available AI tools, provided that they comply with a policy and have been trained?

How to implement AI in your business

To implement AI compliantly and effectively, you should consider the following:

  1. Define use cases and performance criteria: Identify the problem to solve and what success looks like.
  2. Accountability: Assign responsibility for governance and usage and consider new roles or skill sets.
  3. AI literacy: Provide training at appropriate levels to ensure staff understand AI's capabilities and limitations.
  4. AI policy: Develop a corporate policy to guide employees on permitted and prohibited AI use.
  5. Risk assessments: Conduct thorough evaluations before deployment, addressing specific risks for each use case and satisfying legislative requirements.
  6. Operational controls: Implement measures such as AI inventories, safety controls and human oversight.

Take action with expert guidance

AI presents both transformative opportunities and complex risks for businesses. The regulatory landscape is evolving, and responsible implementation is critical. Business leaders should act now to assess their organisation's readiness, define strategic use cases and establish robust governance.

For tailored advice and support in implementing AI compliantly, contact Gowling WLG's AI team. Our experts can guide you through the legal, ethical and operational considerations to ensure your AI strategy delivers value - safely and responsibly.

For more information about our AI team and how we can support you, visit our AI webpage.