 |
| By Chris Graebe |
Lots of folks are anxiously awaiting the next stop-the-presses moment in AI’s maturity.
For many, that means life-altering, jaw-dropping applications that showcase the technology’s potential beyond ChatGPT’s intriguing, yet limited, purposes.
And lots of companies are diligently working on AI applications that can put the glaze back in peoples’ eyes.
While I’m just as interested as the next guy to see what AI can really do, I think there’s a much bigger challenge that requires immediate attention.
That’s especially true as campaigning in an unprecedented and contentious presidential election rages on.
I’m talking specifically about preventing the spread of AI-generated misinformation designed to mess with voters’ opinions and choices.
That’s not paranoia talking. It’s realism. And, thankfully, a growing number of startup companies are making the prevention and identification of misinformation their priority.
Many, not surprisingly, are using good AI to thwart bad AI.
With shares of even the biggest publicly traded AI-related companies pulling back, it might be a good time to consider adding one of these innovators to your portfolio as a long-term investment — because this is hardly a short-term problem.
As though the increase in misinformation across political, health and other sectors alone aren’t enough to attract investors’ attention … new laws and regulations bring the problem into focus even more.
The Digital Services Act in the EU should have a particularly strong impact. The goal of the act is to govern the content practices of social media platforms and address illegal content. It carries fines of up to 6% of a company’s annual global revenue for noncompliance.
Additionally, the Federal Communications Commission outlawed AI-generated robocalls.
And major tech companies have signed an agreement to prevent AI from being used to disrupt democratic elections worldwide.
Now you can see why I think misinformation — and security in general — should take precedence over the creation of the next best AI app.
Just ask Adrian Perkins, who ran for reelection as mayor of Shreveport, Louisiana.
He lost his 2022 bid and blames an AI-created TV commercial paid for by a rival political committee.
It’s believed to be one of the first examples of an AI deepfake used in a political race in the U.S.
The video shows Perkins’ face superimposed on the body of an actor playing him. He’s sitting in the principal’s office. The visit isn’t for cheating on a test or getting in a fight. He scolded Perkins for failing to keep communities safe and create jobs.
Perkins didn’t have the money to fight it and attributed it as one of the reasons he lost the race.
By the way, the ad actually came with a label stating that it was created with “deep learning computer technology.”
Even with full disclosure, the powerful message resonated with and impacted voters.
It happened then, and it’s happening again — many without disclosures.
And this year, it’s feared that the spread of misinformation will reach new levels of sophistication and ugliness with AI in nearly everyone’s pocket.
In a paper recently released, researchers analyzed nearly 136,000 fact-checks dating back to 1995, with the majority of them published after 2016 and ending in November 2023.
They found that AI accounted for very little image-based misinformation until spring of 2023.
Social media sites are the most likely landing place for untruths. It’s the easiest way to tap into more than 5 billion people worldwide.
As you know by now, AI can create videos, pictures and documents that look stunningly authentic.
Misinformation capable of disrupting the election is such a concern that the Department of Homeland Security released a bulletin warning that the technology will likely be used against us in the upcoming election.
These devious acts don’t always look to hurt a candidate’s reputation. The bulletin suggested that enemies might use any number of mediums to inform voters that a polling station is closed or that polling times have changed.
Since a small number of votes can decide an election, keeping a targeted group away from the polls could have a tremendous impact on the result.
Here are …
5 Startups at War with Misinformation
There was a 150% uptick in misinformation in 2023, according to the CEO and co-founder of ActiveFence, a developer of technology that detects malicious content online.
ActiveFence is just one of the startups working to solve — or at least manage — the problem. This New York- and Tel Aviv-based company has raised close to $100 million to date.
Firms working on ways to combat misinformation have certainly caught the eye of venture capitalists as the problem has grown and become more worrisome.
According to Crunchbase, more than $400 million in private equity has poured into the arena over the past couple of years.
Here are a few other startups attracting funding and potentially worth your investment:
Blackbird.ai is a U.S.-based startup that tracks misinformation online campaigns. It uses both human and AI to fight disinformation. This company was the one to notice lots of fake information about 5G cell towers.
Singapore’s Primas hails from and focuses on low-quality content, fake news and plagiarism. It uses blockchain, big data and neuro-linguistic programming (NLP) to find, detect and inform people about any fake or stolen content.
Cyabra is an Israeli startup that protects from disinformation and deep fakes. Its team uses NLP and deep learning. They apply a “truth filter” to help make the digital world a safer place.
Great Britain-based Logically.ai is an app that checks all published articles to identify and fight misinformation or harmful content. It blends together AI, open-source intelligence and a huge fact-checking team.
If you want to pursue an investment in any of the above startups or the many others in this arena, be sure you conduct due diligence before giving up a single penny.
I’ll certainly continue to dig in this area.
Happy hunting,
Chris Graebe
P.S. For either of these groups — those using it for underhanded personal gain and those who are working to stop them — they need data. Lots of it.
And that’s reshaping how data is stored and accessed. Here’s an important story about the companies at the ground level of that effort.
