Friends, today's Cao! story stars Google again — yes, the Google that used to say "Don't be evil."

Here's what happened: Google recently added an AI feature to Google Earth — it lets you generate AI imagery and overlay it directly on real satellite maps. Sounds cool, right? You could add a swimming pool to your neighborhood, put a giant sculpture in front of your office, just for fun.

However, less than 24 hours after the feature launched, Google urgently rolled it back. Why? Because an investigative journalist named Henk van Ess used the feature to generate on real maps:

— A nonexistent nuclear power plant (in a border region)

— A nonexistent refugee camp (in a disputed area)

— A nonexistent fatal car crash (on the streets of Amsterdam)

— A nonexistent bomb crater (next to a field hospital)

Not to be outdone, BBC Verify generated a collapsed Eiffel Tower.

Let that sink in. These things placed on real satellite maps are indistinguishable from reality — especially for people who don't understand AI generation and can't tell AI apart from real satellite imagery. To them, this is "picture proof" — ironclad evidence.

How Many Steps to Turn a 1B-User Map Into a Disinformation Factory?

How many users do Google Earth and Google Maps have? Over 1 billion. Basically, everyone you can think of has used Google Maps. It's the most authoritative, most widely used geographic information platform on the planet — bar none.

Imagine this scenario: right before a national election, a "satellite image" goes viral on social media showing a candidate's mansion hiding helicopters and gold bars. People sharing it say "go check it yourself on Google Earth" — except, wait, the imagery on Google Earth now could have been generated by some random internet user!

Or something more extreme: when tensions are high between two countries, a "satellite image" shows the other side deploying missiles at the border. Do you think that could cause a miscalculation? Do you think someone might take it seriously?

Google's product managers probably thought: "C'mon, people are just playing around — who would take it seriously?" — classic Silicon Valley idealism. They assume all users are well-intentioned, all here for entertainment, no one would use it maliciously. But the reality is: if a tool exists, someone will use it for bad things. And the more authoritative the platform, the greater the damage when it's misused.

The most ironic part? Google has spent recent years on all sorts of projects fighting disinformation and fake news — and then they go and build an AI disinformation tool directly into the world's most authoritative mapping platform. That move is basically left hand fighting right hand.

In the AI Era, What Can We Still Trust?

Honestly, the Google Earth incident is just the beginning. From here on out, AI-generated content will keep getting more, more realistic, and harder to distinguish.

Before, when you saw a photo, you could wonder "is it Photoshopped?" What about from now on? AI-generated satellite imagery, AI-generated video, AI-generated voice, AI-generated documents — every type of information载体 can be faked by AI, and the cost of faking it is practically zero.

So what do we do? We can't just disbelieve everything, right?

To be honest, there's no perfect solution. But a few simple habits can help you avoid 90% of the traps:

First, don't rush to share "breaking news." Especially content that makes you especially angry or shocked — take a moment. Real big news won't suffer from you waiting a few minutes to repost. Wait a few hours, see if authoritative media cover it, see if there's an official response.

Second, cross-verify. If an image says something happened at a certain location, check other map platforms to see if the same location matches. If a story says it was reported by some media outlet, search that outlet's official site to see if it's there. One source might be problematic, but when three to five independent sources all point to the same facts, credibility is much higher.

Third, trust official sources, but with a grain of salt. Government announcements, company websites, public company financial reports — these are relatively credible. But note: what officials say isn't necessarily the whole truth; they might just be selectively telling you part of it. So even with official news, judge it in context and with common sense.

AI has made information production incredibly cheap, and made the cost of辨别 information incredibly expensive. We used to say "seeing is believing" — now seeing doesn't necessarily mean believing. In an era where truth and falsehood are hard to tell apart, maintaining a skeptical spirit might be more important than掌握 any AI tool.

That's all the吐槽 for today. More tomorrow.