Digital Transactions Authoritative, insightful and timely information for payments professionals
Artificial intelligence will be the reason many companies return to the office.
Like every new technology, AI has sparked predictions of sweeping change. But history shows reality is usually more measured. The Internet didn’t erase brick-and-mortar retail or banking. Amazon bought Whole Foods and launched its Go stores, and the United States still has more than 76,000 bank branches.
AI will also reshape business, but in ways more restrained than headlines suggest. Here are my more restrained predictions for how AI will reshape the world of payments.
First, AI will bring employees back to the office because of fraud. As deep fakes become better, people responsible for moving money will want to confirm that requests are legitimate. The easiest way to do this is to walk down the hall and talk to the supervisor in person.
I am not the only one to think this. While giving an anti-fraud presentation and talking about deep fakes, I heard members of my audience talk about moving back to in-person meetings for big purchases. There will be technical solutions to combat AI, but the tech arms race is expensive, and solutions are fallible.
Second, AI will drive more in-person collaboration. Companies will need people onsite to understand how AI affects operations, compare notes, guard against manipulation, and protect the security of internal models and data sets. These systems often run on limited or proprietary data, which can shift quickly with market changes. Having strategists and operators together in the same room will feel faster and safer than relying on remote coordination. This connects to my next prediction.
Third, AI will force us to learn and know more, not less. The current handwringing is about how AI will cause people to rely on AI for everything. But this assumes AI always delivers correct and relevant answers.
In an Instagram post, Sara Landi Tortoli, the founder of Claritys.AI, wrote that two skills are necessary to use AI effectively: prompting, which is asking the right question; and verification, which is understanding whether the output is right, both factually and contextually. For this, a good knowledge base is critical.
A real-world example comes from an interview I did with an AI company, Deep Labs, a few years back. Its fraud-detection algorithms had to turn 180 degrees during the pandemic as consumer and business buying habits changed. In 2019, booking travel looked normal, and buying a lot of electronic equipment for home looked suspicious. But in 2020, when conferences were canceled and people worked from home, it was the opposite.
My last prediction: AI will not eliminate as many jobs as people predict.
Computers can consistently outplay people in chess and even on “Jeopardy!” American Express was using AI for its “Authorizer’s Assistant” when making credit decisions as early as 1988. Nonetheless, we still have credit analysts, and we still have chess tournaments.
There will be jobs where people will be a competitive advantage or even command a premium where customers want to deal with people. AI work is derivative of its training data, so opportunities to stand out in marketing or product design will likely come from humans. Those people may use AI to make their work easier, but that will not be the driving force.
Thinking about AI, I remember reading long ago where someone made a prediction that the Segway would lead to cities being redesigned. Now they are mostly used in another legacy context: shopping malls. AI will have a bigger impact. But even that will be shaped by how people use it.
—Ben Jackson bjackson@ipa.org
