Utilities, Possibilities, and DeepFakes: Can You Spot the Lie?
How Do We Decide What to Believe?
How do we decide to take a political or business action? How do we determine what we believe and what we’ll do? Usually, it’s a function of some research (I hope), what we’re told in conversations, and what we observe in the media. Are things getting better? Worse? Could a conflict be emerging somewhere? Is a company about to make headlines—really good or really bad?
Up until ten years ago, as a society, we had a much easier time identifying good, bad, and malicious information. That’s not to say there were never successful disinformation campaigns to sway markets, push political candidates, or influence geopolitical dynamics. Ah, the good old days.
But today, with the arrival of generative AI, people at all levels of society, education, and job titles are struggling more than ever to understand the information they’re consuming throughout the day—information that forms the basis for crucial decisions about their lives and the society they live in. Should you invest in a given project? Should you transfer a sum of money because your boss told you to on a phone call? Should you sell certain shares because a CEO said something that could tank the company?
The Impact on Business Is Real
AI has a real and profound impact on politics and international relations, but I’ll leave that to my colleagues at The Economist and Foreign Affairs. For business audiences, the risk of disinformation is just as serious.
In early 2023, I was speaking on a panel at the Foro de Emisoras in Xcaret on the topic of AI. We touched on the risks AI poses to markets—particularly in the form of disinformation campaigns aided by AI-generated photos or videos. At the time, it felt far-fetched to many in the audience, who were just beginning to explore AI and for whom generative AI was still very, very new.
By the end of that same year, the Mexican Stock Exchange was hit with a deepfake video of then-CEO Oriol Bosch, spreading misleading information about an investment opportunity. While the video wasn’t nearly as sophisticated as what’s possible now, it was convincing enough to cause a stir, and the Exchange had to move quickly to inform the public that the video was fake.
The Power Is Now in Everyone’s Hands
Today, the ability to produce hyperrealistic videos, photos, and text that seem completely believable is available to just about anyone curious enough to try AI. And while companies like OpenAI are promising to watermark their content and install safety mechanisms in their models, it’s only a matter of time before open-source versions make their way online—accessible to anyone with the skills to download and deploy them.
Voice-cloning models that once required several minutes of recorded speech now need as little as a 15-second sample. Guardrails will always lag behind the current state of play in the AI industry.
Depending on the audience, I share examples of how synthetic identities can be created, refined, and deployed using off-the-shelf AI tools—tools that anyone with an email address and credit card can access. Then I walk through more sophisticated scenarios aimed at targeting companies. Within just 30 minutes, audiences go from “that’s impossible” to “I want to delete my social media accounts and throw away my phone,” quickly followed by “…but then I can’t order an Uber to get home.” It’s that impressive what can be done with AI today—every single day.
Navigating Information in the Digital Age
I’m sharing all this because we now have to do more work for the things that used to take less. In the past, when an article, video, photo, or audio clip was shared with us, we could usually trust it to provide somewhat relevant information—biased, perhaps, depending on the source, but not fabricated.
Now, if we want to succeed, we must seriously consider the source, quality, credibility, and intent of nearly everything we consume. I know it sounds tedious, but it’s absolutely necessary if we want to make sound decisions.
Here are three ways to assess whether a piece of content is misinformation, disinformation, or designed to mislead you using AI. Some of this might sound like common sense—but that’s exactly why we overlook it and make costly mistakes. Not because we’re careless, but because we trusted something without thinking it through.
Where did the information come from? Is it a credible news or data source?
Can you verify it with a quick Google search? If nothing turns up, ask yourself: Are you really that in-the-know that you’ve heard about something before Google has?
Does the video or image have strange shadows, physical inconsistencies, or unnatural movements? Does it make sense that this person would say or do what’s being shown?
Sure, there are many other ways to assess information. But these are simple and useful checkpoints that could keep you from making a rushed decision with real business or financial consequences. The best way to protect yourself is to know what’s possible in the world of AI and information—and be alert about what you’re consuming.
You already do this at the grocery store. You check labels. You make sure what you’re eating is what you intended to buy. You wouldn’t just walk through the aisles tossing in unlabeled items and expect a gourmet meal. And if your partner asked, “Why did you buy this?”—you wouldn’t get away with, “It was at the store, so I figured it was fine.”
In the same way that we make grocery lists to ensure we end up with the meals we want, we should also be intentional about the type and amount of information we take in—especially when making important business, career, and investment decisions.
From Survival to Digital Strategy
At the end of the day, we’re living in one of the most fascinating moments in human history. So yes, it makes sense that we have to work a bit harder to navigate daily life and think about the foundation behind our decisions.
Let me leave you with this: 5,000 years ago, our ancestors spent all day—really, their entire lives—hunting and gathering food and preparing for winter. No breaks, no vacations, and definitely no detours to visit a neighboring tribe (they’d probably steal your food). Today, life’s a lot better. You just need to earn your paycheck—though you probably deserve more—and think a bit more critically about the information you use to get ahead every day. Harder? Maybe. Unreasonable? Definitely not.
Originally published in Spanish for Fast Company Mexico:
https://fastcompany.mx/2024/10/23/mentira-detectar-inteligencia-artificial/
