ChatGPT set records this year for the fastest-growing user base, and when you look what other popular apps such as TikTok and Facebook have done to change our society, it’s hard to wonder what exactly will be the effects of this new AI. OpenAI, a private company, recently announced the release of GPT-4, which is the newest iteration of the logic underpinning ChatGPT, its AI/chatbot technology and ultimately sparked a new rush by tech companies to release their version of generative text chat tools. The major feature of GPT-4 that distinguishes it from prior versions is its ability to incorporate inputs other than text – i.e. photos or images, into searches. Think of Google Lens image search, but on steroids, and that may be how GPT-4 works. Up until now, a user could only interact with ChatGPT by typing questions and receiving written responses back. So now the question is posed: is AI coming for your wealth advisor role and what does it mean for our industry?
How Can GPT-4 Be Used in Financial Services?
Morgan Stanley recently announced that they are piloting the use of GPT-4 within their organization. Is the firm is going to link GPT-4 to their internal records to streamline how their staff looks for investment research and commentary, with the intent to speed up how quickly a user can locate the information they’re seeking with better accuracy? Previously, this would have required going through large amounts of articles and documents manually to find a relevant document. With GPT-4, the AI logic is scanning each article for relevancy and content to deliver desirable results much faster than would otherwise be possible.
Investment research is just one area where financial services can utilize this new tool. Another is automation of meeting organization and notetaking. Right now, a financial services professional meets with a client and then must organize their notes and next actions manually. Instead, GPT-4 may use an uploaded picture of the handwritten notes and automatically move them a client management system and assign tasks and follow-ups to appropriate staff using its logic. For most firms, this is a time-consuming administrative item. We have heard promises of what a virtual assistant can do for years, although up until now most virtual assistants haven’t proven to be very intelligent. This could change with GPT-4. A client meeting may yield tasks like opening new accounts, updating account beneficiaries and changing investment strategies. Even partial automation of these types of tasks would help firms scale up their operations without having to increase their headcount.
AI will make it easier for small financial services firms or solo practitioners to compete with their larger competitors. Previously, extra resources and scale required more staff, but GPT-4 may change that if utilized intelligently. GPT-4 may be able to identify market trends and signals to be acted upon quickly by active money managers or correlate investment research at scale.
Potential Challenges with GPT-4
One immediate issue of concern is data security and client confidentiality. There will be a big temptation to use the technology to automate things like emails and meeting notes. Unfortunately, that requires copying and pasting sensitive client data into GPT-4. Things are moving so quickly that there hasn’t been any real serious discussion around how client information is protected. Since GPT-4 is now available online to anyone with a web browser, it’s also hard to limit a user’s access to the app, which means we must rely on everyone being prudent with their usage of it. With cybersecurity and client data security increasing rapidly in importance, financial services firms need to be careful what checks and balances are in place for their usage of GPT-4.
Additionally, there are already numerous stories about GPT-4 returning odd responses to queries and questions, meaning it is still unrefined and users need to quality-check the output they’re getting. Even Google and Microsoft will freely admit the technology isn’t perfect and the responses need to be checked for accuracy. Since the major tech firms are now in a race to come out ahead in AI, they are moving forward as fast as possible with their offerings, meaning there will be growing pains with this technology. There is a huge temptation in financial services to use this technology to save time, but it requires care and diligence.
Expertise Is Required to Keep Pace
Financial services firms that want to be competitive and participate in any technological gains will need to dedicate time and resources to staying on top of all the changes and capabilities new AI tools are bringing. While Microsoft and Google are currently making waves in this area, there will be other entrants racing to catch up. Players like Amazon, IBM, Adobe and others with less familiar brands will be competing as well. Time will tell which companies have developed the best methods for training AI and how they perform in their dedicated tasks. Financial services firms will very likely have many platforms to test and choose from soon, but custom programming and implementation sometimes mean high switching costs. Technology teams in our industry will have their work cut for them for years to come.
Shane Cummings, CFP, is the Director of Technology/Cybersecurity at Halbert Hargrove, where he also works with clients as a Wealth Advisor.