AI Governance Cyber Risk and Leadership Transformation

Lloyds Bank welcomes AI into the boardroom

AI is turning up in new situations and on new frontiers. With the vast investment in the last year into data centres, chips and new AI models, there’s all kinds of strange new developments happening in the world of AI.

This week, footage emerged of a Polish-speaking robot chasing wild boars off the streets of Warsaw. In finance, Lloyds Bank looks set to become the first major UK business to bring AI into the boardroom. The bank will use a custom-built AI agent from the UK advisory business Board Intelligence, which was founded by Pippa Begg. While many organisations are wary of using AI at decision-making level due to the biases in AI tools, this tool aims to focus on eliminating or at least drawing attention to human biases. “Begg said the agent could help to eliminate ‘human bias’ from board decisions, including on making acquisitions, and analyse executive and employee performance,” writes The Times. “Very few people have access to two things that are important for being able to deliver this,” said Begg. “One, an understanding of what typically doesn’t go so well in executive rooms and boardrooms, and therefore what ‘good’ looks like – because there’s a relatively small population of people that sit in those rooms. And then the second piece is to deploy technology in a secure way. Most organisations don’t want to give their internal teams access to some information. Our set-up enables us to securely host all that information without others being able to read it.”

“For now, Lloyds is mainly making Board Intelligence’s bot available to bosses to prepare for meetings. Begg described this as ‘step one’ for most organisations, where AI is used to ‘augment the human’s ability to consume information, and form judgments and test their judgments and test their theories, before they get into the room’. The second step, she said, will involve board members having an open laptop in meetings and ‘being able to almost interrupt and say: ‘Hang on, I think you’re falling into this trap.’ Or: ‘I disagree.’ Begg added that allowing AI to actually have a legal vote could be a ‘dangerous leap’.”

Welcoming our robot overlords

Chief executives at banks tend to be pressed for time. Mark Zuckerberg may soon offer a solution: the virtual chief executive, able to interact with everyone in the company, day and night. Mr Zuckerberg has been accused at times of being socially robotic, but soon he won’t be able to escape that charge. Zuckerberg has been working on a solution, apparently spending hours every week helping to code new AI systems and sitting in on technical reviews, according to news stories.  “The Meta chief is personally involved in training and testing his animated AI, which could offer conversation and feedback to employees, according to one
person,” the FT reports. “They added that the character is being trained on the billionaire’s mannerisms, tone and publicly available statements, as well as his own recent thinking on company strategies, so that employees might feel more connected to the founder through interactions with it. The Zuckerberg character will be trained on images of the chief executive as well as his voice, one person said. If the experiment is a success, influencers and creators might one day be able to do the same, the person added.”

Meta employees are being encouraged to use company tools and the open-source software OpenClaw to design AI agents that will make their jobs more efficient. It’s not really clear if Zuckerberg’s virtual boss will work everywhere. Tech companies have fundamentally different business models from banks, and are often led by their founders rather than experienced chief executives that have come up through the bank.

When the security company is too good at its job

UK regulators are the latest to review Anthropic’s new AI model Mythos after the model proved capable of finding IT and security loopholes in bank security systems – and being able to create ways to exploit the loopholes. “Leading British banks, insurers and exchanges will be warned about the cyber security risks exposed by Anthropic’s latest model, Claude Mythos Preview, at a meeting with the regulators in the next fortnight, according to two people briefed on the talks,” the FT reports. Regulators fear that such software would be exploited by bad actors looking to steal from financial institutions and threaten financial stability. Anthropic recently created Project Glasswing with 40 partners including Amazon, Apple, Microsoft and Cisco to test its new AI model. “During our testing, we found that Mythos Preview is capable of identifying and then exploiting zero-day vulnerabilities in every major operating system and every major web browser when directed by a user to do so,” wrote Anthropic in a blog post. “The vulnerabilities it finds are often subtle or difficult to detect. Many of them are ten or twenty years old, with the oldest we have found so far being a now-patched 27-year-old bug in OpenBSD – a system known primarily for its security.” US Treasury Secretary Scott Bessent had called in US bank CEOs last week to warn them about the risks from the new model. All of the major bank CEOs, who were already in Washington for a Financial Forum, attended the meeting apart from Jamie Dimon. This is all dream marketing for Anthropic, which said it will hold back on a general release of the model – though company leader Dario Amodei said he expects open-source companies and Chinese developers to develop similar models within a year.