Do you have any artificial intelligence (AI) secrets? Maybe the better question is: Would it concern you if your most intimate chatbot queries were shared with your family, colleagues or bosses?
For all the hype, AI still carries a stigma. I have been tracking the mounting backlash for a while. It has become deeply uncool to admit that you used AI to help you with your art, not to mention that it will be judged more harshly. And trust in Big Tech leaders, often for valid reasons, is abysmal. Bashing AI on the internet gets a lot more applause than defending it.
Part of the reason for this secrecy is that even as business leaders encourage adoption, research shows that it’s still ostracized in the workplace. More Americans now talk openly about their AI use with managers and colleagues, a June study from Atlassian found, yet doing so carries a “steep penalty.”
Even when the output itself was identical, participants rated hypothetical peers who disclosed using AI as 10 times “lazier” and were 24 percentage points less likely to recommend them for a high-visibility project. No wonder employees are unlikely to shout their prompts from the rooftops.
This may also start to explain one of AI’s ongoing economic conundrums. Productivity gains not visible in macro data may be accrued by individual workers who have little incentive to reveal them.
But this culture of secrecy can create other risks for businesses, whether it’s data leaking to unapproved tools, or a lack of knowledge-sharing across teams. The obvious first step is a clear, accessible policy spelling out when AI can be used and when it must be disclosed. Managers should also encourage prompt-sharing, tips and open conversations about experiments.
There are other consequences to a tech shift unfurling in the shadows that extend well beyond the office. Ask friends how they use chatbots when no one is watching, and the answers might make you gasp.
A digital archive project dubbed ‘What We Tell AI’ is already collecting anonymous confessions via boxes placed in public spaces across the US. They offer a glimpse at the increasingly intimate ways this technology is being embraced.
People are using AI to compose birthday messages to relatives or cheat on homework and the New York Times Spelling Bee game. Others admit they are in love with ChatGPT, or view it as a best friend.
They are not the use-cases that dominate earnings calls.
AI firms publish reports on how people are adopting their products, but some researchers say this is necessarily selective. Firms preparing for major public offerings or defending enormous spending may be incentivized to highlight productivity gains and have less reason to dwell on other uses.
That leaves policymakers with an incomplete picture of a technology proliferating rapidly in people’s jobs, relationships and private lives, including children’s. Regulators hoping to not repeat past mistakes should demand greater transparency from tech companies to let external researchers study how their products are actually being utilized and what effects they are having.
The secrecy also reveals a larger problem for the industry. They are still left hoping that eventually its value will be stronger than the shame of turning to their services. That contradiction is especially stark among youth. Young adults in the US are the most likely to think AI will have a negative effect on society, but they are also the group that says they use chatbots the most, according to a June Pew survey.
The AI revolution is not being televised; it may be too disjointed for that. The message that this technology could put us all out of work, or at worst kill us all, is hard to dismiss as paranoia. But it is not stopping people from asking for dating advice either. That may tell us more about AI’s future impact on society than another survey on how many minutes it saves on digital coding tasks. What people are unwilling to admit matters more than what the AI industry is willing to measure. ©Bloomberg
The author is a Bloomberg Opinion columnist covering Asia tech.