Clara Durodié, CEO of Cognitive Finance Group, a self-described “pragmatic techno optimist” and keynote speaker at the 2023 Finance Awards, which took place in Mondorf-les-Bains on the evening of 21 November, began her talk by noting that she had completed her speech for the event over three weeks ago. But “ever-evolving news coming from the AI sector has compelled me to revise the speech week after week,” she said. And with the corporate governance shakeups at OpenAI and Microsoft over the past weekend, she had to write her speech yet again.
“AI regulations have been pouring down, and that’s a good thing,” Durodié argued. “The European Union AI Act is on its last innings; the UK hosted the first AI Safety Global Summit to start the global process of understanding AI risks, and the United States laid down the framework of AI regulations, while China has led important discussions and policies around the governance of technology.”
But “the pace of change is truly, truly unprecedented,” she highlighted. Durodié compared the rapid developments in the field of artificial intelligence to being on a high-speed train, where the landscape is a hazy blur--you can see cows, forests, a river, but you can’t actually make out the details because the train is moving so quickly. “Decision-makers face the daunting challenge of finding the strength to slow things down and understand what is actually happening.”
The world is at a “pivotal juncture,” she said. The “geopolitics of AI and its future are being shaped as we speak.”
Risk oversight and governance are key
Though she’s “endlessly captivated” by the possibilities of artificial intelligence, Durodié’s background as an investment management business executive means that her focus throughout her pivot into artificial intelligence has remained on AI risk oversight and governance, instead of “mindlessly chasing new technologies and the latest shiny new thing.”
AI is an amplifier. It has the force to magnify what it finds in the data--and this alone brings significant risks
“AI is an amplifier. It has the force to magnify what it finds in the data--and this alone brings significant risks. Therefore, extra considerations are essential.”
AI for business growth, not cost-cutting
“Contrary to popular belief, AI is not about increasing productivity and cost-cutting,” Durodié argued, tackling some of the myths around artificial intelligence.
“The primary objective of artificial intelligence is business growth. We use this new technology to find new opportunities and keep growing our reach. Productivity is just the byproduct of business growth--it’s not the ultimate objective. And as for cost-cutting, that is a finite area. You cannot keep cutting costs; otherwise, you have no business left.”
“Another myth--which you have heard many times--is the misconception that AI serves as the magical potion, a solution to all the challenges, including economic growth.” This is not the case, and she “urged caution” against grand predictions and astronomically high numbers for growth. The fundamentals--infrastructure like roads or trains, for instance--need to be fixed before expecting AI to drive economic growth.
Building and maintaining people’s trust in artificial intelligence is essential, Durodié added. “Without trust, nothing of value will be accomplished with this technology.”
Don’t rush hysterically into AI deployment
The “fear of missing out” on the benefits of artificial intelligence is a danger as well, with the pressure on boards or CEOs often resembling “hysteria” rather than innovation. “It has led to hasty AI deployment decisions since ChatGPT was launched earlier this year. According to a recent McKinsey survey, 82% of business leaders deploying generative AI cite internal pressures as the primary driver of that change.”
Organisations should favour impartial advisors with no vested interest or corporate conflicts, technology-savvy advisors that will help the board prioritise long-term strategic thinking over short-term irrationality
“I advocate that organisations should favour impartial advisors with no vested interest or corporate conflicts, technology-savvy advisors that will help the board prioritise long-term strategic thinking over short-term irrationality.”
There’s also a “crucial need” for AI literacy amongst board and business leaders, argued Durodié. Business leaders need to understand artificial intelligence in order to make “informed” decisions for their firms.
“Rushing business leaders to act for immediate adoption needs careful consideration,” she said. “This is no time to rush. Rushing it now for those who have done not a lot to date on AI may require the counter intuitive decision to slow it down.”
“Huge responsibility” of the finance sector
The 500-some attendees at the event were players active in Luxembourg’s financial sector, coming from banks, law firms, insurance companies, asset managers, private equity firms and the political world. And this means they have a “huge responsibility,” argued Durodié.
We are not in the business of selling shoes. We manage money
Technology errors in the industry can have significant consequences. “We are not in the business of selling shoes. We manage money,” she said. “Rushing deployment without thorough consideration may erode customer trust, undermine company resilience. Supervisory board members must refrain from casual remarks about AI replacing staff, without a solid understanding of the technology’s intricacies and limitations.”
Be cautious of over-reliance on AI
“The final myth I want to unpack is over-reliance on AI,” said Durodié. “We have to be very careful.” Current AI technology prioritises optimisation. But too much emphasis on optimisation may, for instance, lead to AI models posing drastic measures like exterminating humans to optimise waste management in a large city, she said, offering an extreme example.
“That’s not good. Optimisation with AI is widely misunderstood, and this is reinforcing the need for a mandatory switch-off button.”
Trends for 2024
2023 will likely be remembered as the year when artificial intelligence entered the “collective consciousness,” argued Durodié. What does 2024 hold?
The explainability and transparency of AI will be an important trend to look out for next year. With increased reliance on complex algorithms, it will be essential to have transparency regarding these algorithms. “The ability to understand and explain AI decisions becomes very important, especially in our sector. This transparency builds trust and shows responsible use,” she said. “When AI hallucinates--meaning it’s making things up--the regulators are not pleased. And they shouldn’t be.”
“With our regulations and responsibilities, I put it to you that deployment of generative AI is a fantasy suitable only for creating new AI companies claiming to fix these hallucinations. So why would you bring hallucinating systems in your company just to pay other companies to fix them?” asked Durodié. “Addressing [hallucinations] is very difficult, very complex and very expensive. But by all means, use generative AI. But do it safely and do not rely on it.”
We will see--it’s very likely--more activity not only to create laws, but also to apply the ones that already exist
AI regulations will be a second big trend, she argued. “We will see--it’s very likely--more activity not only to create laws, but also to apply the ones that already exist. Indeed, in the United States, the chair of the SEC went on-record to raise the bar of AI regulations in financial services to mitigate what he identified as predictable crises.” In addition, “regulators will likely expect companies to self-regulate.”
On a broader level, Durodié said that she expects to see an evolution towards smaller AI models, which could allow new ways of data analysis and decision-making that are “perhaps better, and perhaps cheaper.”
“Hyper-personalisation is another trend. As customers’ expectations evolve, so does the demand for hyper-personalised services,” she argued, including in the financial sector. “And quantum computing will emerge--and is actually emerging--as a game changer in processing vast amounts of financial data at unprecedented speeds.”
“Be responsible stewards”
Durodié concluded by emphasising that AI adoption is not simply about understanding the technology and using it to grow a business. “It’s also about adopting a mindset,” she said.
But “as we navigate this very fast-moving landscape, I urge you to be more than adopters. Be pioneers, be inquisitive thinkers, critically reason over things that are being said to you about this technology. But also be responsible stewards of this technological frontier--not only for us, but also for your children. We are shaping their future.”
And how can the investment management industry can shape the future? By “being intentional about what you invest in.”