2 January 2024

ChatGPT one year on: Where can AI go from here?

Can AI really get us to a “Minority Report” Future? Recent developments suggest so

The past two years saw the widespread proliferation of AI. The deployment of humanoid robots to everyday tasks, including managing and interacting with customers, is not as futuristic as many people may believe. Image: iStock.

By Ran Mo

In the 2002 film “Minority Report”, a special police unit can arrest murderers before they commit their crimes. Fiction or soon-to-be fact? Don’t underestimate the possibility of it being the later.

With the rapid popularity of ChatGPT — the large language model jumped to a million users just five days after its launch in November 2022; can we anticipate that artificial intelligence (AI) will truly make a difference for us in the future, especially the banking and finance space, which is ripe for further disruption? Based on events of this year, the technology looks set to both make and prompt changes.

Indeed, in the dynamic landscape of asset management the integration of AI is transforming the way data is analysed, decisions are made, and information is disseminated. BlackRock’s upcoming rollout of generative AI tools to clients is just one example of the industry’s increasing reliance on AI-driven solutions. While the technology offers immense potential, we can’t help but ask: what kind of era are we in; where will AI take us; and is all of this ultimately just our beautiful imagination?

Surface and the depths of financial data

Traditional financial data analysis has largely been focused on structured information such as financial statements and transaction records. However, a significant portion of valuable insights lies beneath the surface in unstructured text, including news and social media information. The emergence of research AIs marks a new phase in financial data analysis, allowing investors to tap into the hidden depths of information.

We’ve heard this story: During Francis deSouza’s last earnings call as CEO of a gene sequencing company Illumina, he faced challenges related to the contentious US$8 billion takeover of Grail. By analysing deSouza’s speech using AI, researchers noted that there were shifts in his speech patterns, increased filler words, and audible signs of tension when discussing Grail, which was a sign for anxiety when talking about this sensitive issue, even though he told the analysts that the drama was only affecting “a very small part of the company”.

Less than two months later, deSouza resigned. Considering some high-profile fraud cases in the US like the management of Enron, Madoff, and Theranos, for example; it’s evident that they all maintained a facade of innocence in front of the media and investors before their fraudulent activities were exposed. If AI technology had been available during these times, however, each of the aforementioned companies wouldn’t have been able to deceive the public for so long.

Quite simply, AI can provide and analyse more comprehensive information, and like many investors, we sincerely hope for significant advancements in the field.

Biases and decision-making challenges

The discussion of adopting AI in the investment industry is not new. Since the effectiveness of AI systems fundamentally relies on the quality of the data they receive, if datasets are incomplete or unrepresentative, it may restrict the objectivity of AI, and biases within the development teams training these systems can inadvertently sustain and perpetuate such biased cycles.

Achieving complete objectivity, aligned with a manager’s preferences, remains a difficult task. The need to ensure fairness in decision-making processes is a critical consideration in deploying AI in asset management.

These considerations are a given, in light of both the rewards and risks associated with deploying technology to make an investment decision, which in turn could determine huge gains or equally as large losses. But how about AI’s application in financial communications?

Privacy dilemma and ethical considerations

In the investor relations space for instance, AI can assist professionals in creating profiles and accurate portraits of investors based on behavioural data. It can then use intelligent matching algorithms to push relevant information to individuals and institutions that are most likely to make investments. Moreover, AI can provide specialised and real-time services to each shareholder, a task traditionally considered valuable but challenging task.

There could be a catch, however: privacy. This dilemma surrounding AI stems from its insatiable need for personal data, raising concerns about storage, usage, and access. Questions about data origin, storage location, and unauthorised dissemination challenge traditional data protection laws. The evolving AI landscape underscores the urgency for ethical guidelines to minimise privacy risks.

For communication professionals more broadly, another issue with AI is copyright. The known strength of AI lies in its ability to provide powerful solutions, heavily reliant on high-quality training data. Many countries’ copyright laws require AI developers to obtain permission from copyright holders before using their content for training. Some AI developers, seeking to avoid lengthy negotiations and time constraints, resort to using pirated content for training. This practice poses challenges for copyright holders, as the backbox nature of the training process makes it difficult to provide evidence of infringement.

In response to concerns about OpenAI using unlicensed content for AI tools, some media like the Guardian has blocked the use of its content by OpenAI.


Then there’s “AI-washing”; a term analogous to the concept of greenwashing. Even in the communications industry, there are firms claiming to leverage AI to make mission-critical decisions without providing clients as well as staff of proof of this being the case.

To that end, companies are being warned against making exaggerated claims about their use of AI. The US Federal Trade Commission (FTC) emphasises the importance of substantiated claims, cautioning against overuse and abuse of AI as a marketing term.

In short, the scrutiny on false or unsubstantiated claims highlights the need for transparency and accuracy in portraying AI capabilities.

Balancing potential and responsibility

We can see that perhaps there is still a journey ahead before fully realising AI in the service of humanity. No wonder many people say AI is merely a game for the ‘picks and shovels’ providers.

However, just as I felt fascinated by cryptocurrency and chose to be in Fintech when went back to business school a few years ago, I believe AI will change the world. To ensure responsible and effective AI integration, I hope that the governments will implement more feasible regulations and stakeholders must remain vigilant, fostering an environment where the power of AI is harnessed responsibly and for the benefit of all.

Strategy, Technology, Uncategorized