Beyond the hype: A practical insight into the implementation of artificial intelligence models as part of London Tech Week 2024

1. AI can help make companies run more efficiently and boost human creativity 

From the creative industries to cybersecurity and beyond, AI has the potential to improve efficiency and unlock human creativity by freeing up time to concentrate on more highly creative tasks. 

Gemma Jennings, Product Management Lead at Google Deepmind highlighted that large language models (LLMs), such as  Google Gemini, have the benefit of working at scale which means they have the ability to improve and optimise people’s day-to-day work. Tino Millar, Founder & CEO of Move.ai, which creates 3D motion capture content from 2D video with AI, explained how video game and film production companies are benefiting from AI.

Moreover, in the cybersecurity sector, the adoption of generative (Gen) AI and LLM, and training those models around specific behaviours, can transform analysts’ ability to rapidly “observe, orient and act”.  For example, James Todd, CTO at Adarma explained that if a cyber security company is observing an IP address or asset identifier, historically this would mean having to undertake the labour-intensive task of going back to configuration management databases or HR records to determine who that was and to assess the value of that asset to the organisation. Combining the implementation of AI alongside a  ‘human in the loop’ approach, allows companies to accelerate this process – utilising the best of human and tech skills – to significantly reduce the time it takes to analyse and respond to a cyber threat. 

4. Collectively consider risks

Exploring and understanding the risks around the introduction of AI is also critical for successful adoption. Developing a risk framework unique to each company’s individual needs, and educating employees to ensure they understand the positive attributes – and risks – of using AI is crucial for businesses looking to introduce AI to their operations.

James Todd highlighted that organisations need to understand the associated risks around implementation and the underlying capability that businesses require, including the risk brought by models and algorithms created internally, in addition to the inherent risks and vulnerabilities that can occur through the adoption of AI. 

The panel agreed that companies need to take a cross-organisational and joined-up approach to addressing risk. Gemma Jennings explained that at Google, responsibility is shared across teams – from product managers through to research scientists and ethics team members and beyond. This collaborative approach gives the organisation the chance to pressure test innovation and stretch its thinking. 

Adding to ideas around collaboration, Tino Millar noted that companies could learn from the creative world, developing a sandbox where teams can take an iterative approach to AI development and implementation, allowing them to safely explore big gains and value for the company and protect from unwanted risk.

Digital Security by Design (DSbD) is a government-backed initiative to transform technology and create a resilient, and secure foundation for a safer digital future. Through collaboration between Digital Catapult, researchers, industry and government, these new capabilities will free the way business and people use and trust technology. Sign up to the DSbD newsletter to stay up to date with upcoming events, news, insights and opportunities.