Brad Smith, Microsoft's vice chair and president, is frustrated. After 33 years at the company, he's watching the tech industry damage its own credibility by catastrophizing about artificial intelligence's impact on jobs, and in doing so, alienating precisely the young workers who should embrace it.
"Nobody knows for sure, but let's not panic," Smith said from Microsoft's Redmond headquarters, summing up a message he believes his peers desperately need to hear.
The complaint runs deeper than rhetoric. Smith argues that tech leaders have become hypocritical and grandiose in their warnings, undercutting what he sees as a transformative opportunity for workers. When Anthropic published an essay calling for a global pause on AI development, Smith's response was blunt: if you genuinely think the technology is moving too fast, slow your own operations down rather than lecturing others.
The real damage, Smith contends, is being done to recent graduates. Young Americans who came of age during COVID, spent formative years on screens, and navigated political turbulence are now entering the workforce. Instead of being presented with opportunities, they're hearing apocalyptic narratives.
"This is being presented to them as something that is going to happen to them, not for them," Smith said. The reaction from graduates booing AI at this spring's commencement ceremonies should serve as a wake-up call to the entire sector.
Smith identifies several specific failures in how technology leaders have communicated about AI and employment.
The timeline problem looms largest. AI's economic transformation will unfold over 25 years, not 2.5 years. But entrepreneurs raising capital have incentives to accelerate narratives. A two-decade disruption period is actually faster than historical precedent for complete economic shifts, yet the compression serves fundraising purposes rather than accuracy.
Tech leaders also systematically overestimate technology's immediate impact while underestimating human adaptability. "There was a time when humanity discovered that a horse could run faster than a person," Smith noted. "So people learned how to ride horses." The same principle applies to AI. Instead of replacement anxiety, the framing should center on augmentation.
A third pattern troubles Smith: confidence divorced from evidence. "You find that the same folks who made the wrong predictions a decade ago keep making them with extraordinary conviction," he said. Bad forecasts from the past decade are recycled with undiminished certainty, providing easy material for news coverage and social media engagement.
The calls for regulation suffer from similar hollow grandstanding. When some companies demanded legislation on social media in previous years, they simultaneously opposed every specific bill in Congress. Smith sees the same dynamic threatening to repeat with AI policy, with proposals so sweeping they'll never survive legislative scrutiny.
The path forward, Smith believes, requires reframing the conversation entirely. Technology should be discussed not as a threat to neutralize but as a tool to amplify human capability. The industry needs to listen to young workers' skepticism rather than dismiss it, and prove through action that AI development benefits workers rather than replaces them.
For Smith, the core message is simple: stop predicting doom and start demonstrating how technology makes jobs better. The generation entering the workforce is watching closely enough to know the difference.
Author James Rodriguez: "Smith's frustration is justified, but tech's credibility problem wasn't built in a day and won't be fixed by exhortations alone. Actions matter more than messaging now."
Comments