How to Tackle the Dangers of Mainstream Deepfakes

It is curious (and maybe profound) to find an article impacting cognitive security and disinformation issues in the business section of the newspaper, but The Washington Post included an examination of the technology behind deepfake generators on the front page of the Business section on March 28, 2021. Anyone who spends time doom scrolling on social media will have seen the humorous videos of dead celebrities rocking out to some anachronistic song, so maybe it should be neither curious nor surprising that much of the technology that fuels our day is driven more by lucrative info-tainment opportunities rather than any more profound sense of the moral, ethical, or legal considerations. New apps like those discussed in “Anyone with an iPhone can now make deepfakes. We aren’t ready for what happens next,” are flooding the mainstream with simple technology to manipulate photos and videos. Like most technology advances, we are compelled to create before we fully consider the implications of creation.

The suggestions of developing some type of “code of conduct” for developers, reinforced by technical countermeasures like digital watermarking, are worth more detailed consideration on the practical implications of implementation. Will the private sector willingly go along, or will state, federal, or international regulation be needed? Can we incentivize businesses to do this on their own to protect their reputations and intellectual property, like we’ve done for other digital IP in the past (and will that be sufficient)? These sorts of questions are best addressed before the technology proliferates too widely. In the early days of the war in Iraq, insurgents were skillful in funding and broadcasting real videos that they would take out of context and use for their own needs. U.S. forces struggled to respond and correct the record after the information was out there; once you tap into readers’ deep-seated biases, it becomes a high hurdle to change their minds. At the time, insurgents were using much less sophisticated means, and yet our forces remained reactive and on the defensive the entire time of our involvement there. Without more tools in the toolbox to address and counter deepfakes at the speed and scale needed for future conflicts, we will be increasingly vulnerable to the kinds of dissension and confusion our adversaries aim to seed.