WAIFC members have requested a deeper and broader focus on the implications of artificial intelligence for the future functioning and offer of financial services.1 To start a series of opinions, the writer asked AI itself how it view this: “What does ChatGPT think the impact of AI will be on the financial services sector?”.
The first half of this opinion is the writer’s unresearched impressions on the subject in March 2025. The second part of this opinion, written in italics, is the answer ChatGPT gave.
What The Writer Thought
There is no on/off switch for AI, or computerization or digitization. The progression of these technologies has been coming upon us for the better part of a century as a rising tide on the beach, with each wave pushing a bit further up the shore.
The writer is, admittedly, a follower and not a leader when it comes to the
adoption of AI. The first signs of language-based generative AI came as uninvited guests when writing texts on the cell phone and in Word documents or equivalent Google texts. I do not remember exactly when it started, but my phone began to make suggestions that may – or may not – complete my thoughts, and to correct my grammar. The prompts imposed themselves on the screen, unpleasantly, and were impossible to ignore. Yet, overall, with practice and determination, these AI assistants might be helpful boosts for productivity in daily office life.2 They do make a person think about whether the prompt is the right way to go, or not.
For years now, calling financial service firms has required passing through an
initial automatic voice filter with choices and verbal instructions to guide the query along. These can be irritating, too, as when a person’s accent in a language is not recognized and there is no way around that blockage. When logging into a financial firm online, too often an artificial face invites interaction in a virtual space to the right side of the screen and asks all kinds of questions to figure out the problem. The company is imposing its new and “fun” (read “cheap”) way of working through enquiries. This handling of client enquiries may not be nurturing for a commercial relationship. It is hard to distinguish the effects of any one utterly transformative step in the process of moving first to computers – then online – and then online with AI features, the evolution experienced in large parts of the world since the 1970s.
In financial services, algorithmic portfolio management and trading of securities and futures has been a part of capital markets and portfolio management for years at this stage, first launched by the largest firms with the greatest resources to get a technological advance on their less well-financed, smaller peers.3 Again, there was no light switch moment. Financial management and trading position limits set by algorithms are not at first generative AI, because the computer-based algorithms are written by a person. But as the algorithm “learns” and “improves,” then another step has been taken.
Going back further in time to measure the distance travelled,4 the first public, regulated exchange trading based on computers arrived in November 1977 at the Toronto Stock Exchange with the launch for brokers of the CATS software (Computer Assisted Trading System). Based on desktop computers at brokers’ offices instead of chalk boards and open outcry orders on a trading floor, CATS was not intelligent. What it brought was transparency and an organized way to trace and respect the time-price priority of orders in an electronic queue. These were significant steps. Floor traders and portfolio managers everywhere looked on with heightened interest – and also with great concern about their future jobs.
In the following years, portfolio managers and securities traders got computerized, massively so, eventually working with powerful desktop set-ups with multiple screens, and when out of office with hand-held tablets. As foreseen from the moment CATS was introduced, there was job rotation and not everyone adapted to the constantly changing IT environment; but overall, there are more humans working in financial services than before the first computerization of the industry. The office may be unrecognizable, but the people are there and they are accomplishing more in an industry that is many times larger.
At the turn of the millennium, we heard that “dot.com” companies were a paradigm shift, businesses were being created the likes of which the world had never seen. The crash that followed left damage that was, in fact, all too familiar. Is AI going to be a change the likes of which the world has never seen? Or will more pervasive usage of AI in our industry come to our desks as more digestible tools that can be managed by humans working more efficiently, more “intelligently.”
What seems evident in early 2025 is how easy it has become to search for information, to get comparisons or quickly verify facts when something comes to mind. (Those quick searches definitely need verification.) Where might the next changes come from as broad language generative AI moves further into everything? This should facilitate comparison shopping for financial services of all kinds, the obstacle being that corporate information is not presented in ways that are relatively easy to compare side
by side. We will see how generative AI handles the poor comparability question, and also whether companies redo their digital presences to be more or less visible for searches. The reverse is going to be even more true than we have seen so far – corporate outreach to customers whose data have been sold off and who then become prefiltered, inexact new customer targets.
It seems best to use – not to lose - our heads when it comes to AI, and to remember that somewhere in all this there are people.
What Chat GPT Answered
AI in financial services is evolving rapidly, and by late 2025, we can expect significant developments across several key areas:
1. AI-Powered Personalization & Customer Experience
Banks, insurers, and asset managers will refine AI-driven personalization, offering hyper-customized financial products and advisory services based on real-time customer behavior and predictive analytics.
2. Regulatory Technology (RegTech) & Compliance
AI will play an even bigger role in monitoring transactions for fraud, automating compliance checks, and detecting financial crimes with real-time risk assessments. Expect increased collaboration between AI developers and regulators to ensure ethical, transparent AI usage.
3. AI-Driven Investment Strategies
Quantitative trading and robo-advisors will integrate deeper AI capabilities, using alternative data sources (e.g., satellite imagery, ESG signals, social media sentiment) to refine investment strategies.
4. Generative AI in Banking & Insurance
Generative AI will be widely used for document automation, contract analysis, fraud detection, and personalized financial reporting. Large institutions may deploy proprietary AI models fine-tuned for specific financial tasks.
5. Decentralized Finance (DeFi) & AI
AI will enhance DeFi by optimizing lending protocols, risk assessments, and fraud prevention in blockchain-based financial systems. Expect increasing AI-powered smart contract auditing tools.
6. AI & Sustainability-Linked Finance
With growing ESG (Environmental, Social, Governance) concerns, AI will help financial institutions measure and report sustainability risks, analyze green investments, and ensure compliance with carbon regulations.
7. Ethical AI & Bias Reduction
By late 2025, AI in finance will be under greater scrutiny for fairness and bias reduction. Companies will need to demonstrate AI decision-making transparency, particularly in credit scoring and lending.
Would you like insights on any specific AI applications in finance?
1-WAIFC opinions begin with a visual image of the subject, in order to set the scene. ChatGPT actually has no visual image of Itself – the writer asked it the question of what it looks like. No response. Instead, the image for this opinion is a simple question mark found in a quick online search.
2-It is unsettling to think that the software is trying to learn about me and my interests in order to “serve” me better.
3-ChatGPT does not have an answer as to when this practice first began; the author did ask.
4-The ARPANet set up by the Pentagon connected the first computers in 1970, but this was not yet the internet. The World Wide Web was created by Tim Berners-Lee and launched at CERN in Switzerland in 1989.