Amplify or magnify - the AI trust gap

Every day I visit LinkedIn or some other social media platform, I see two types of posts emerging: people who have embraced AI wholeheartedly and become LLM evangelists, and those who either have or think they’re losing their jobs to AI, and are convinced it’s a threat to humanity on par with the splitting of the atom.

Both sides have depressingly cogent arguments. My position falls somewhere in the middle.

In the time I’ve been in the working world, organisations have increasingly become consumed in what I call busywork. Layers of bureaucracy and hierarchy can transform a simple task into weeks of navigating office politics and trying to read the tea leaves. It feels exhausting, occasionally demeaning, but most of us have learned to grumble and accept it.

 Until that is, generative AI came along around late 2022. It’s been less than three years and like Facebook in its heyday, suddenly it’s all anyone is talking about. Companies are sinking massive amounts of money into it (remember how every company – from startups to Wall Street, just simply couldn’t do without social media managers?), and those who position themselves as early adopters are in hot demand. Like desperate lovers seeking psychic readings, companies are hiring AI influencers to not only review but in some cases completely redesign their business and operating models to be AI-first.

 Something must have gotten lost in the memo, because polls and surveys worldwide show that humans increasingly distrust AI. Pope Leo XIV was the latest to join the debate, publishing a 43,000 word essay pleading with the world to address the threats posed by AI (and there are many, starting with data centres that consume more drinkable water than entire cities).

 As an example, a recent global survey of 23 countries by EY shows that Australians have one of the most negative views of AI in the world. Seventy-seven percent of Australians report that they use AI. They trust it in low risk settings, such as customer and retail services, but not in areas like government and financial services. That’s what we call a significant trust gap. So, what we can do about it?

 While minds more lofty than mine can make much more eloquent arguments for or against, let me posit my humble case – that the problem with AI lies in both how it can amplify or magnify. The best use case for AI in the workplace lies in how it can amplify and enhance people’s existing skills and capabilities. Going back to my point earlier about busywork – it can do this by removing the administrative burden that weighs down a lot of teams. Spreadsheets and templates that used to take a couple of people and a manager days to complete, review and sign off can suddenly be turned around in eight hours. By one person.

 An overworked and over-committed senior executive can upload a bunch of fragmented thoughts into their AI of choice, and it can turn it into ten different types of content (which sound frighteningly similar once you detect the pattern – but that’s for another blog post on another day). But if you have a 5am and 9pm meeting across timezones on the same day, a conference to attend in between, stakeholders to schmooze and urgent work from your team waiting to be signed off – all you want is something that can organise your chaotic thoughts into coherence. That is the golden promise of AI. That it can do so not only a billion times faster, but much more efficiently and cheaply.

 The downside to this? AI can also magnify mistakes. Without human input and judgment at crucial points of the chain, the dominoes fall over one by one. Think of the written word – now undervalued because AI can churn out paragraphs of prose faster than writers can think. Your average AI can create reports, presentations, white papers, strategies and crunch through data analytics before you have time to finish making a cup of tea. Amazing and inspirational, until on closer read, you discover that it’s made several not only very serious, but potentially litigious mistakes. But couched in such beautifully business-like language, sprinkled through with all the right jargon, that it makes total sense if you don’t think about it much.

 Amplify or magnify. That is the question. And the answer is essentially human. We still live in a human world, and AI is a tool built (or so we’re told) to assist humanity. Let’s keep it and ourselves accountable and instead of replacing, we should be looking at amplifying.

All content is copyright © Karen Tay. Unauthorised reproduction, scraping, or use of this content, including by AI training systems, is prohibited without express written permission from Karen Tay.