The 5 Biggest Threats of Artificial Intelligence for Humans

Right now, there are AI programs that can pass the bar exam.1 If you don’t know what to do with that information, you’re not alone.

What is the future going to look like as these tools become even more advanced and more widespread? Will our future be bright, or will we descend into some nightmarish dystopian nightmare? Will AI save us or doom us?

Like Jeff Goldblum’s character says in Jurassic Park, “Scientists were so preoccupied with whether they could, they didn’t stop to think if they should.” Let’s consider that sentiment here and examine some of the biggest risks of AI.

1. Hallucinations

No, we’re not talking about a Hunter S. Thompson drug trip, we’re talking about generative AI passing off incorrect information as if it were real — a concept industry insiders refer to as “hallucinations.”2 Generative AI models are really good at sounding convincing, even if what they’re telling you isn’t real. In a world where the Truth — capital “T” — is tenuous at best and fact-checking is an all-but-dead art, hallucinating AI could lead to further obfuscation of what is real and what is simply a dream of electric sheep.

One particular area of concern is online content. Let’s say you google, “how to change the oil in a Toyota Tacoma.” How do you know the first page that pops up was written by a human who knows what they’re talking about and not an AI program that hallucinated something vaguely related to engine oil? The degradation of available media is a massive concern if AI starts generating the majority of it.

2. Cyber Security Concerns

AI tools are getting pretty good at coding, which is both good and bad news for developers everywhere.3 It also means the barrier to entry for creating malicious programs just got a lot lower. If one were so inclined, it would be pretty easy to use an AI tool to write a program that could do untold damage, like stealing identities. (With that in mind, you may want to read up on our favorite identity theft protection services. In this age, they’re well worth the investment.)

Smart consumers need to see the writing on the wall and start shoring up their cyber security right now. We don’t know what new threats are over the horizon, but rest assured they’re there.

3. Deepfakes Blurring the Lines of Reality

Remember when that picture of Donald Trump scuffling with the police went viral? How many people do you think stopped to investigate its validity before they shared it?

We’ll settle it now: No, the former president did not get into a brawl with the NYPD. The images were what’s known as deepfakes, a classification of AI-generated images, videos, speech, or other media that mimic reality. They’re getting so good that experts have no idea how to regulate them.4

Similar to hallucinations, we need to understand how to grapple with deepfakes before they further blur the line between reality and nonsense.

4. Upending the Labor Market

When GhatGPT started hitting the mainstream, one of the first questions on a lot of people’s minds was, “How will this affect my job?” The truth is, no one really knows. Investment firm Goldman Sachs estimated that AI could replace the equivalent of 300 million full-time jobs,5 but the impact could be much more widespread.

So widespread, in fact, that some experts are calling on policy-makers to seriously consider introducing some form of universal basic income to offset the effects of millions of people suddenly being out of work.6

5. Roko’s Basilisk

We admit that this one is a little far-fetched, but it’s worth bringing up. Roko’s basilisk is a thought experiment that posits that the development of AI will eventually lead to an all-powerful future intelligence that will have the ability to create a simulation — that we’re currently living in, for the sake of the exercise — and use it to retroactively punish anyone who didn’t help bring it into existence.8 It could have happened already, and may be occurring right now. 2020 was pretty hard, right? Was that the Basilisk?

We’re (kind of) mentioning this tongue-in-cheek, but there is a real risk that unchecked AI could become so developed that it decides humanity is no longer needed and wipes us all out.9

Final Thoughts on the Threats of AI

That was a bit of a bleak read, but he good news is there are companies actively working to create guardrails around these tools to prevent some of their worst tendencies10— from telling us the wrong information to getting into our defense systems and launching all of the world’s nukes at once.

Hopefully AI will ultimately make all of our lives easier. If used correctly, these tools could take some of the drudgery out of our days, shorten our work weeks, and maybe give us more time to invest in our more human passions.

If we want to reach that reality, though, we have to start thinking about the best ways to limit what these tools can and can’t do. We have to remember the wise words of Dr. Malcolm. We have to think about if we should, not if we could.