Google’s AI Overviews Can’t Seem to Answer If 1995 Was 30 Years Ago

Was 1995 30 Years Ago? Google's AI Overviews Struggles with a Simple Math Question

In 2025, artificial intelligence often grabs headlines for both impressive feats and notable missteps. While AI is reshaping industries and influencing daily life, it still stumbles on straightforward tasks—like simple arithmetic.

Google's AI Overviews, a tool designed to provide quick summaries and answers, recently showed signs of confusion over a basic question: "Was 1995 30 years ago?" Mathematically, the answer is clear—subtract 1995 from 2025, and you get 30. Yet, Google's AI sometimes responded with 29 years or even contradictory answers within the same reply.

What’s Going On With Google's AI Overviews?

Users quickly noticed inconsistencies and shared screenshots online showing Google giving incorrect responses to this simple query. At times, the AI answered correctly, but other times it faltered, delivering mixed or wrong results.

This inconsistency highlights a broader issue: AI tools, despite their sophistication, are prone to errors—especially with tasks that seem trivial to humans. The problem isn’t just with math; previous incidents include fabricating meanings for made-up idioms and getting facts wrong about sports rules.

Why Does This Matter for IT and Development Professionals?

• Trust but verify: AI-generated content and answers can be helpful, but they aren’t infallible. Always cross-check critical information.
• Expect imperfections: Even leading AI tools can produce errors, so integrating them into workflows requires oversight.
• Stay informed: Following updates and known issues with AI platforms can help you adapt and avoid pitfalls.

For those working with AI or relying on it for research, this serves as a reminder to maintain skepticism and not accept AI output at face value.

Where to Learn More About AI Tools and Their Limitations

If you want to deepen your understanding of AI capabilities and how to work effectively with these technologies, consider exploring specialized courses that cover AI fundamentals, prompt engineering, and automation. Resources like Complete AI Training's latest AI courses offer practical guidance designed for professionals.

Until AI tools reach higher reliability, a simple calculator or mental math remains the best way to verify if 1995 was indeed 30 years ago.