A nonprofit digital library has assembled a collection of old AI-related media — ads, toys, television clips, and sci-fi ephemera — from the decades before generative AI took over the news cycle. The project frames itself as a nostalgic look at artificial intelligence before the topic became loaded with anxiety about jobs, copyright, and misinformation.
The archive itself isn't a business tool and won't change how you run your company tomorrow. But the timing is worth noting, because it lands during a period when AI marketing fatigue is becoming its own story. Surveys of business buyers increasingly show skepticism toward vendors slapping "AI-powered" on features that used to just be called automation or search.
This isn't the first AI hype cycle, and the archive is really a reminder of that fact. Artificial intelligence as a marketing term dates back to the 1950s, and it has gone through at least two well-documented boom-and-bust periods — the 1970s and late 1980s — often called "AI winters," when funding and enthusiasm collapsed after promised breakthroughs didn't materialize on schedule. Expert systems, natural-language interfaces, and even early chatbots were all sold with confidence that, in hindsight, outran the technology.
What's different this time is scale and speed. Generative AI tools reached hundreds of millions of users within a couple of years, and the underlying technology — large language models — has demonstrated genuine, measurable capability gains, not just clever demos. That's a real distinction from earlier cycles, where the gap between demo and deployable product was often wide and permanent.
Still, the pattern of vendors overstating what their product actually does hasn't gone away — it's just wearing newer branding. That's the throughline connecting a 1980s home robot toy to a 2026 SaaS dashboard promising "AI-driven insights."
For small businesses, the practical takeaway isn't nostalgia — it's due diligence. When a vendor pitches an AI feature, ask what specific task it automates, what data it was trained or tested on, and what happens when it's wrong. Those questions cut through marketing language regardless of which decade it's coming from.
History also offers a cost lesson. Previous AI winters were triggered partly by businesses and governments overinvesting in systems that couldn't deliver, then pulling back hard. Something similar is already visible in enterprise software: several high-profile AI pilots at large companies have been quietly shelved over the past year after failing to show return on investment. Small businesses with limited budgets have less room to absorb a similar miscalculation, which argues for smaller, reversible bets — a monthly subscription you can cancel, not a multi-year platform contract — until a tool proves its value on your own numbers.
There's also a customer-facing angle. As AI backlash grows — visible in surveys where consumers say they distrust AI-generated content or dislike chatbot customer service — businesses that lean too hard on "AI" as a selling point risk alienating some customers rather than impressing them. How you describe a tool to your team is different from how you should describe it to your customers.
Watch for two signals over the next year: whether major AI vendors start pulling back marketing language in favor of specific, measurable claims, and whether enterprise AI spending data (tracked by firms like Gartner and IDC) shows continued growth or the kind of pullback that preceded past AI winters. Both would tell you where we are in the current cycle.
The bottom line: AI hype has a long history of running ahead of AI capability, even though today's tools are demonstrably more capable than their predecessors. Evaluate vendor claims on specific, testable outcomes rather than the label attached to the product, and keep AI commitments small and cancelable until they've proven their worth in your own operation.