AI Outlook Summer 2025: practitioner insights

Vik Khurana has recently been quoted in Thomson Reuters Practical Law's article “AI Outlook Summer 2025: practitioner insights”. 

The article explores the key artificial intelligence (AI) issues currently impacting their clients and the developments they anticipate in the near future. It highlights that navigating the regulatory framework of the EU AI Act ((EU) 2024/1689) is a primary focus for many businesses, particularly regarding risk classification, compliance timelines for general-purpose AI (GPAI) models, and AI literacy training. You can read the article in full here (subscription required).

Some interesting insights include Vik's comment on the impact of the European Commission's delays in adopting codes of practice and harmonised standards and the lack of clarifying guidance. He said: 

"particular pain points include translating the Act's requirements into language the development teams can understand and reflect into their products, given the lack of harmonised technical standards which were promised by the European Commission when it launched the Act".

He went on to discuss data protection issues relating to the development and deployment of AI tools, notably ensuring that clients have 

“the proper legal basis, controls and protections in place to process personal data, whether in the training of AI systems or in the operation and use of those systems which could ingest or output information relating to data subjects”

When asked about the rise of agentic AI, Vik said: 

"On the technology front, we're closely following the rise of Agentic AI systems, which promise to redesign end to end processes and workflows. While there will be substantial benefits to businesses, this comes with different types of risks to LLMs, caused by unique features of Agentic AI such as reduced human oversight, accountability and control, the higher risk of inaccuracies at each step of a business process, and data, privacy and cybersecurity issues when agents operate across multiple data systems and applications. We are helping clients review their AI governance to ensure they are scalable for Agentic AI."

Read the entire article here (subscription required).