A common assumption about deepfakes:

Is it so unthinkable that a government would—fearing deepfakes and perhaps lacking a robust censorship system like China’s—regulate access to hardware? Add graphics cards to the list of “restricted” exports, like dangerous chemicals?

Let's play around with this lifeworld a little bit more.

“A graphics card? In France, you’ll need a license for that.”

Paris, 2023:

Fearing deepfakes, France requires customers to purchase a license to buy graphics cards with certain specifications. Lawmakers wave off the EFF as extremists. (One senator’s remarks are translated into English as “ideological colonialists of the American variety,” to much mouth-frothing on Fox News). Lawmakers construct the laws using analogies to gun licensing.

Meanwhile, Ledger, a niche domestic hardware company, sees an opening: they start manufacturing graphics cards that meet France's legal specification. The law goes on the books in 2024, production starts October 1, 2023. They’ve already signed an exclusive deal with Sony to supply parts for the EU model of the next Playstation.

Imagining panicked policy

For deepfakes, there are a few reasons why requiring licenses for computer hardware might sound appealing as far as quick fixes go:

1. On the supply side, it's easier to control hardware than software. You can shut down manufacturers and monitor imports. Software (and other web content) is slipperier, safer to deliver across borders over encrypted channels.

2. On the demand side, it’s more difficult for citizens to evade a hardware ban than a software ban. To get a physical item, you’ve got to leave your house and meet with someone face-to-face. The barriers to entry are higher.

3. Domestic manufacturers may benefit from the arrangement. As the scenario above illustrates, local companies can use the regulations to capture marketshare away from foreign competitors whose products are non-compliant.

While regulating access to hardware is probably not a good idea, it does sound like something that someone might try. (Compare to encryption debates in the UK and US). Yet the possibility is conspicuously absent from prognostications about the near future of cybersecurity. This omission is especially bewildering given a brewing conflict over microchips between the United States and China.

It’s our job to consider these possibilities. If we want to protect societies against deepfakes, we need to think about how we should deal with them and how people might choose to deal with them in practice.

Panicking regulators could create panicked policies: quick “fixes” that make it look like someone is doing something. And policymakers’ decisions have infrastructural consequences. (We’re still living with decisions Congress made about the radio spectrum in 1910s).

An exercise for you

Actors: Here I thought about policymakers in Western Europe who can’t easily haul US tech CEOs in front of their legislatures. Who else might get caught up in these decision-making processes? Tech CEOs? Legislators in Taiwan? Intelligence organizations? Activist groups? Who else?

Technologies: Here I thought about deepfakes. They're in the news; they cohere to debates around disinformation, truth breakdown. What other technologies might trigger panicked policy? One technology that pops into my mind: cryptocurrency mining, which speaks to debates around climate change and wealth inequality. What else?