AI Needs Babysitters, and These Stocks Are Cashing In

Key Points

• Google Translate has been around since 2006 and still can't handle the nuances that matter in business documents.

• Fiverr's fastest-growing business segment involves humans teaching other humans how to manage AI tools.

• Even the most advanced AI tools get trained on human-created content and need constant human supervision to function properly.

• 10 stocks we like better than Fiverr International ›

Some people fear that artificial intelligence (AI) systems will replace the online tools you're using today.

• Why would you look for human freelancers on Fiverr (NYSE: FVRR) or Upwork (NASDAQ: UPWK) when you can just ask ChatGPT to write a web page?

  Where to invest $1,000 right now? Our analyst team just revealed what they believe are the 10 best stocks to buy right now. Continue »

• Google Search is great and all, but Alphabet (NASDAQ: GOOG) (NASDAQ: GOOGL) could surely deliver better answers to your questions with the Gemini large language model (LLM).

• And human translators must be halfway to the unemployment office already, right? With automated translation services like Google Translate, you only have to click a button to make that legal document or user's guide available in Greek, Japanese, and Slovenian. The work is done by a neural network, auto-trained on public web data. Duolingo (NASDAQ: DUOL) should just shut down because nobody needs to practice actual language skills anymore. AI can just do it for you.

Those are the market-moving worries hanging over stocks like Duolingo and Fiverr. I get it -- change is scary.

It's not that simple, though. AI can help the very companies that seem to be under robotic attack.

Experts suggest (and I agree) that the machines still need lots of human hand-holding. As a result, one of Fiverr's best-performing businesses these days is humans providing AI management and instruction-writing services. Alphabet built a Gemini-based info-snippet tool into Google Search, and its system developers are always hard at work optimizing the underlying search algorithms.

Image source: Getty Images.

Even AI needs a human touch, according to Fiverr

Speaking at an industry conference last week, Fiverr CEO Micha Kaufman highlighted the need for humans in the AI industry.

"You need very high-skilled people that know how to use AI," he said. "That portion of our business has been growing very nicely. You see some verticals mostly around technology, and marketing, workflow, agent development and so forth, growing 10x."

Everybody wants to make the most of AI-based resources, Kaufman explained, but the machines absolutely require human handlers. AI platforms can do some impressive things these days, but the best results always involve a human touch along the way.

My Swedish perspective on AI's limits

The translation stuff is kind of personal to me. I've spent half my life in Florida but was born and raised in Sweden. I've seen moose and northern lights in the wild, you know. Translation is a side gig for me, adding a little pocket money and some variety to my daily routine.

AI tools are helpful but unbelievably far from perfect.

Even when the source and target languages are closely related, like my favorite pairing of Swedish and English, the best machine translation tools sound robotic and make lots of mistakes. There is often no "right" word or phrase but a choice of near-synonyms with subtle differences. The nuances can make a big difference in medical or legal documents, or in detailed instructions for using a high-tech gadget.

The choices you make at these incredibly common forks in the road will steer other sentences in specific directions. Any language is a creative system. Switching from one to another is a genuinely human effort. The machines just can't get the all-important nuances right -- yet.

The translation services that connect me to gigs rely on several layers of translation (possibly machine-assisted) and human reviews -- often done by several different translators. Some of their clients (my bosses' bosses) insist on human translation and will pay extra for a better result.

I mean, even the AI tools rely on original translations by human experts. It's a lot like ChatGPT or Gemini getting trained on large heaps of hand-typed documents. The neural networks need great examples of what a Swede would say in a particular situation, for instance.

The translation industry agrees: Robots need review

Falk Gottlob, chief product officer at the Smartcat translation platform, expects high-powered LLMs to become commodities over time. People and companies will prefer accuracy and trusted data flows over raw processing muscle.

"Generic translation engines trained on internet data may look powerful, but they consistently fail to meet enterprise realities like brand standards, compliance, and cultural nuance," Gottlob said in an emailed analysis. "Based on our work with global enterprises at Smartcat, we've found that the most strategic approach integrates human expertise directly into AI workflows."

Sure, you can run a fully automated workflow at Smartcat, simply pulling direct translation results from service providers like Google Translate or Microsoft (NASDAQ: MSFT) Translator. But nobody wants that. Without human review, the paying client pretty much always ends up with a stiff, error-riddled, and practically useless translation. The real value in Smartcat (where "cat" stands for computer-aided translation) isn't its ability to deliver plain machine results but in facilitating the human analysis that follows.

I know things will get better over time, but this is not a new problem. Google Translate, for instance, is one of the best tools for this purpose, and it has been around since 2006. And it's still not nearly good enough.

AI helpers are just lifeless tools in human hands

When I look at AI tools like ChatGPT, Gemini, and the neural network versions of Google Translate, I see exactly that -- tools, best used by a skilled artisan.

That's why translators still have jobs, adding value to Duolingo's services. It's why Fiverr and Upwork make money by helping humans help the machines. All of this may change if I ever get my hands on a human-like artificial general intelligence (AGI), but that's probably many decades away.

So I don't mind investing in Alphabet, Fiverr, or Duolingo while other investors see huge AI risks hovering around them. Stock discounts based on misunderstandings are always welcome.

Tack så mycket, börsen! (That's my spin on "Thank you very much, Wall Street!" in Swedish, but Google Translate won't quite agree.)

Should you invest $1,000 in Fiverr International right now?

Before you buy stock in Fiverr International, consider this:

The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Fiverr International wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.

Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $648,369!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,089,583!*

Now, it’s worth noting Stock Advisor’s total average return is 1,060% — a market-crushing outperformance compared to 189% for the S&P 500. Don’t miss out on the latest top 10 list, available when you join Stock Advisor.

See the 10 stocks »

*Stock Advisor returns as of September 15, 2025

Anders Bylund has positions in Alphabet, Duolingo, and Fiverr International. The Motley Fool has positions in and recommends Alphabet, Fiverr International, and Microsoft. The Motley Fool recommends Duolingo and recommends the following options: long January 2026 $395 calls on Microsoft and short January 2026 $405 calls on Microsoft. The Motley Fool has a disclosure policy.