Less than a week after Donald Trump took office, in January 2017, I reported a story that scared the crap out of me. By that point, it was already clear to most of the world (with the exception of Mark Zuckerberg and, possibly, Trump) that fake news had played a significant role in our country’s division and Trump’s election. To wit, a 28-year-old man named Edgar Maddison Welch had recently opened fire inside a D.C. pizzeria because he had read apocryphal journalism claiming that Hillary Clinton was running a pedophile ring in the basement. As bad as the fake news stories that dominated the presidential election were, I noted, they were essentially a beta version of what was to come. With new developments in machine learning, the fabrications of 2016 would pale in comparison to what we would see and hear in the midterms and in 2020, once anyone could use so-called deepfake technology to mash up, manipulate, and inject fake videos into the social-media bloodstream.

Ever since, dozens of news articles and research papers have described a future in which we struggle to distinguish simulation from reality. Some have warned of the peril when people stop using deepfakes to put Kim Kardashian’s face on an anonymous porn star’s body, say, and start creating synthetic video of Kim Jong Un discussing his nuclear arsenal. Or fake audio of Vladimir Putin and Trump’s private meeting, where any number of fake scenarios could be floated, comrade. Even the most amateurish efforts could have election-bending effects. Last week, for example, a conservative news outlet, CRTV, spliced together an interview with Alexandria Ocasio-Cortez, which had originally aired on PBS, to make Cortez look like a bumbling fool. When CRTV published the video on Facebook, it received over a million views in 24 hours. As the Verge noted, most of the people who watched the clip “appear convinced it’s real,” based on their comments and shares. (The video is still on Facebook, now with over 3.3 million views, but has a little text update at the bottom of the page to say that it is “satire.” Sure.) Earlier this month, Robert Chesney, of the University of Texas School of Law, and Danielle Keats Citron, of Yale University, published a paper warning that the worst is still to come with these fake videos. In the paper, “Deep Fakes: A Looming Challenge for Privacy, Democracy, and National Security,” the pair reported that the ability to distort reality has taken an exponential leap forward and that machine learning and A.I. are rapidly escalating the technology’s sophistication, thereby “making deep fakes ever more realistic and increasingly resistant to detection.”

On Monday, online pranksters decided to go after Cemex, a Mexican cement company, attempting to discredit the company by posting fake videos on YouTube depicting, of all people, Tom Hanks and Steven Spielberg as pedophiles who ran a child sex ring for the company. (The creators were evidently inspired by the Comet Pizza example.) As Ben Collins, a reporter for NBC News, tweeted, “Pandemic levels of bullshit unchecked on YouTube today.”

On some primal level, fake news was an inevitable result of malicious actors gaining the power to turn online mischief into real-world mayhem. What’s most exasperating, however, is that the biggest tech companies—Facebook, Twitter, Alphabet’s Google and YouTube—allow it to go forward with relatively few checks. In fact, these companies have done next to nothing to stop the spread of this kind of fake content, even though it’s all people have been able to talk about for the past two years. When Recode’s Kara Swisher asked Mark Zuckerberg earlier this month why content like this lives on Facebook, he defended Holocaust-deniers’ right to spread misinformation by suggesting it was an expression of their naïveté. It was a glib, foolish response. And anyone who knows Zuck, a no-nonsense cutthroat businessman who built a $511 billion company ($630 billion a week ago), knows he couldn’t have possibly believed a word of his own bullshit.

In reality, Silicon Valley listens to two masters: financial incentives and regulation. Technology companies have largely ignored fake news because, until now, they didn’t have to pay attention to it. There were no consequences for policing their platforms. Sure, the media issued harangues, but their stocks were peaking. And Snap, which briefly tried to model itself as a safe haven, appeared to be cratering; many users just didn’t care. But then came the first murmurs of a double-barreled reckoning. Last week we saw tech stocks plummet (and continue to do so) after Facebook reported that users had started to abandon its services, and that ad dollars seemed likely to follow in the coming quarter. On Monday, Democratic Senator Mark Warner released a paper outlining a vision for legislating these big tech companies. It delineated 20 different options for addressing the industry’s unchecked power, from micro-level solutions to sweeping policy changes. It called for combating disinformation, protecting user privacy, promoting competition, and combatting fake news. “Due to Section 230 of the Communications Decent Act, Internet intermediates like social-media platforms are immunized from state torts and criminal liability,” the paper notes. “However, the rise of technology like DeepFakes . . . is poised to usher in an unprecedented wave of false and defamatory content.” While the paper acknowledges that changing Section 230 would meet with dissent from tech providers and digital-liberties organizations, it suggests that there is a problem here that needs to be fixed before things spiral even further out of control.

If Warner’s recommendations had come out weeks ago, I think they would have been glossed over by tech executives, possibly as a media-hungry gesture from a canny senator who may have larger political aspirations. But after the massive Facebook-led tech-stock implosion, which erased tens of billions in market capitalization for these giant companies, many observers now believe that Wall Street has a limit to what it’s willing to tolerate. As one investor in the social-media space told me on Monday: “After Zuckerberg testified to Congress and the Facebook crisis hit its peak, the stock actually shot up—obviously that was a vote that people didn’t care. That was proven wrong last Thursday, during Facebook’s earnings announcement, to the tune of $120 billion in market-cap loss. The takeaway here is we best be careful if we’re going to underestimate public opinion about the aloof, ambivalent, and callous ways these companies have been handing these very serious issues.” The investor added: “The stock going up showed a shrug of the shoulders, yet that proved to be a dire mistake.”

In business, as in life, timing is everything. Had companies like Facebook and YouTube done more months or years ago to stop this, we wouldn’t find ourselves racing against technologies that are now being used in the real world and will, once again, likely affect the outcome of an election. But the ability of these companies to police themselves, or the government to reprimand them in a timely manner, was never so seriously. This week, Congress and the president are talking about 3-D-printed guns like it’s something they just discovered, yet people (myself included) have been warning about what was to come for the past decade. For Trump to turn around this week and say that this technology “doesn’t seem to make much sense!” is a testament to how backwards our government really is.

But what happens when we fast-forward five years from now? When the battle between truth and lies on these social platforms will be so intense—with advanced machine learning, artificial intelligence, and new types of video content that hasn’t yet been invented—that eradicating propaganda and made-up content might prove impossible? If Senator Warner does have larger political ambitions, or if he really wants to stop the spread of fake news, he’ll need to do more than just release a paper about what could be done to rein in these tech giants that now have a chokehold on how we all get our information. He should start pressing for legislation that will force these companies to act. As the past two years have shown, if we leave it up to people like Mark Zuckerberg to do on their own, it’s only going to get worse, not better.