The House Intelligence Committee heard alarming testimony Thursday that deepfake videos could be weaponized by foreign adversaries to sow divisions in the United States.
Clint Watts, a former FBI special agent and senior fellow for Alliance for Securing Democracy at the German Marshall Fund, warned lawmakers that Russia and China will likely both work to develop “synthetic media capabilities” for use against the U.S. and other adversaries.
“China’s artificial intelligence capabilities rival the U.S., are powered by enormous data troves to include vast amounts of information stolen from the U.S., and the country has already shown a propensity to employ synthetic media in television broadcast journalism,” he said.
{mosads}Chairman Adam Schiff (D-Calif.) described the videos as a “nightmarish scenario” to legislate.
He also called on social media companies to take action, adding that waiting until after the 2020 elections would be too late.
Republicans also voiced concerns about deepfake videos during the hearing but expressed fears that filters to control the videos could treat conservatives unfairly.
Rep. Devin Nunes (Calif.), the top Republican on the committee, said the filters, if left in the hands of tech companies, would be created by left-wing partisans.
“How do you put in filters to these tech oligarch companies?” Nunes asked at one point in the hearing. “There are only a few of them … that are not developed by partisan left-wing,” he continued, stating that “most of the time it is conservatives who get banned and not Democrats.”
Rep. Will Hurd (R-Texas), a former CIA officer, told reporters the government should do “basic research” on how to combat deepfakes that can be shared with both agencies and the private sector.
He said social media companies should work on their ability to identify the creators of deepfakes, while the State Department or FBI should focus on prosecuting actors linked to foreign countries.
Rep. Jim Himes (D-Conn.) cautioned that the government will have to avoid over-reaching when it comes to policing deepfake videos, warning parody or satiric videos should not be censored.
“I think it’s a really hard question, but I would really urge and will urge the Congress to proceed very, very carefully, because as scary as deepfake technology is, the prospect of damaging our rights to free expression, the right to satirize politicians is also pretty scary,” Himes told reporters.
A fake video of Speaker Nancy Pelosi (D-Calif.) last month that had edited audio to make it appear she was drunk has intensified concerns, particularly among Democrats, of the deepfakes threat.
Rep. Peter Welch (D-Vt.) told The Hill that there are “categories of harm” to both politicians and consumers that could be raised by deepfake videos, while describing the current regulations around social media companies as “the Wild West.” Social media platforms should look into creating “a standard of reasonable care” for their users, he said.
Another committee member, Rep. André Carson (D-Ind.), told reporters that he is currently researching legislation that would protect elected officials against deepfake videos, adding that lawmakers having to defend themselves against this type of threat is “unacceptable.”
One major topic of conversation at the hearing was the need to educate the public about how to spot deepfakes.
Steve Grobman, the senior vice president and chief technology officer of cybersecurity group McAfee, told The Hill in a phone call that the public’s ability to identify deepfakes needs to get to the level of identifying photoshopped pictures.
“At the end of the day we need to work off of the premise that the public is going to be exposed to deepfakes, and the public needs to recognize that the credibility of the source is as critical as the content they are viewing,” Grobman said.
Grobman added that the public needs to have “a healthy level of skepticism” toward any videos of politicians that seem crafted to influence their views of that person.
But experts also say deepfake technology is growing increasingly more sophisticated and readily accessible on the internet — to the point where it will fall on forensic experts to discern whether content is real or fake.
And they say as the technology becomes more advanced, so too will the technology to detect deepfakes — it will be a never ending race.
“I don’t know if we can ever be completely prepared for this,” Schiff said. “But there is a lot more that we can do to both identify our enemies’ plans and intentions, ferret out what they do as early as possible, expose it, establish a deterrent, and inform the public about what they need to look out for and do so in a way that they don’t simply disregard anything and everything they see.”