News You Can Use

The UK Government have published their AI Opportunities Action Plan, their strategy on how to leverage AI as a nation. This plan covers 5 core themes: boosting economic growth, ethical AI, skills development, infrastructure investment, and public trust and collaboration. The overarching desire is to use AI to boost the UK economy and support struggling industries, such as healthcare. There is also an interesting proposal for an National Data Library to support AI training.

Increased adoption of AI in the public and private sector could see an increase and potential normalisation of AI in legal. This could result in an increase in queries from our clients regarding AI use but also internally from our lawyers. This advice could range from the practical ways to get the most out of AI and support in getting started, to an increase in compliance queries. This relates to our AI workshop offering and, with AG currently working with various public bodies from local councils, government, and the NHS, there may be an opportunity for us to support this public sector evolution.

Marc Benioff, CEO of Salesforce, shared insights on how AI is reshaping industries by driving productivity, enabling automation, and transforming workflows. He emphasised the growing importance of AI-driven tools like Einstein GPT in Salesforce's ecosystem and highlighted that businesses must focus on reskilling their workforce to thrive in the AI-driven era. Benioff also stressed the ethical responsibility of leaders to ensure AI is used responsibly, particularly in protecting jobs while fostering innovation.

Benioff's views align with AG's approach to leveraging AI tools responsibly while preparing employees for changes in workflows. For AG, this means enhancing internal AI training initiatives (like AGPT onboarding) to empower employees with skills to maximise AI tools' potential. It also reinforces our responsibility to balance AI deployment with safeguarding ethical and employment considerations, ensuring our adoption of tools aligns with client trust and firm values. However, it indicates that there is definitely a move from some in the market to realign resource internally by relying on AI, we have to be conscious when speaking to clients that priorities will differ.

Spellbook have released their new Playbook feature to the wider market now, with a concerted push into in-house teams. It allows you to draft a series of rules to assess an open document against, with the ability to add optional drafting, explanations, comments to insert. Performance varies depending on complexity of the concept you are trying to assess, with prompting being really important to get this right.

This shows another vendor who previously focused on law firms moving to look at the in-house market. The Playbook feature is impressive on simple docs and may enable clients to self-service on some work, but struggles with performance on some aspects of complex legal work. We have Spellbook and can arrange a demo / look into this for those interested. This is a concept that clients are super impressed with (the ability to markup docs in Word) and even in its current form we have clients asking whether they should purchase.

Sam Altman, CEO of OpenAI, reflects on recent progress in AI, including advancements like GPT-4 and multimodal capabilities, while acknowledging challenges such as scaling, safety, and aligning AI systems with human intent. In contrast, Gary Marcus critiques AI's lack of common sense reasoning, a problem persisting despite decades of AI research. He highlights the need for AI systems to move beyond mere pattern recognition to exhibit a deeper understanding of context, causality, and human behaviour. Marcus proposes exploring hybrid approaches combining symbolic AI with deep learning to bridge these gaps.

Both discussions underline the opportunities and limitations of current AI tools. For AG, Marcus's critique is particularly relevant as we explore AI applications for legal analysis, where contextual understanding and logical reasoning are critical. Leveraging hybrid approaches or developing domain-specific frameworks could help us mitigate gaps in generative AI's reasoning capabilities. Altman's focus on safety and scaling reinforces the importance of embedding robust testing and compliance checks into our AI tools to ensure reliability in use cases like contract analysis.