Endless Infowar

At a time when US politics is generally believed to be highly partisan, it might come as a surprise that the Senate Intelligence Committee has managed to reach apparent unanimity on the issue of Russian online disinformation. The conclusion of the independent reports it issued this week is that there was a lot of it, both during and after the 2016 Presidential election, and that it benefited the Republicans. That one of the reports, from Oxford University, emphasised that the chief goal was to foment polarisation makes this bipartisanship all the more ironic. The headline claim is that Russia targeted African-Americans in 2016 with a view to encouraging abstention. One reason why the Republicans are happy to subscribe to this is that it shifts the focus away from their efforts at systematic voter suppression. The impact of social media disinformation is unpredictable and probably trivial: there's no shortage of evidence that African-Americans were disenchanted with the Democrats after the failures of the Obama years and many hadn't forgiven Hillary Clinton for her "super predators" remark. On the other hand, the impact of voter suppression through gerrymandering and impediments to registration and the ballot box is proven, both in terms of its effectiveness and its increased scale in recent years.

Another reason for the bipartisan consensus is that the American national security establishment, which encompasses both parties, wants to see an increase in funding for counter-measures and "proactive engagement" - i.e. US online disinformation. Just as we once had a "missile gap" (which turned out to be fake news, incidentally) so now it is variously suggested that the West is playing catch-up in respect of disinformation, that it is at a disadvantage to Russia in "asymmetric warfare", and that it is particularly vulnerable to new technology like AI-enabled deep-fake video. All three are specious claims. The US has been a pioneer in disinformation and covert propaganda since the 1940s and it has been leveraging its home advantage in the Internet since the 1970s, something that was made all too clear by Edward Snowden's revelations. The idea that disinformation is a "weapon of the weak", and that the US has paradoxically suffered due to its conventional military strength, is a recycling of excuses first minted in Vietnam. The claim that deep fake videos will lead to "a world in which there is no truth and no trust" is the most ideologically telling, combining as it does a suspicion of technology with a contempt for the gullibility of the masses, both of which are traditional features of conservative thought.

The emerging panic over "deep fakes" ignores that the technology to produce convincing moving images has been around for over a century and has taught us two things. First, that truth doesn't collapse, largely because most people understand the nature of illusion and quickly adjust to new forms of it on exposure (the tale of viewers terrified by the Lumiere brothers' oncoming train is likely an urban myth). Second, that the technology to identify fake imagery has advanced in lock-step since the move from video to digital CGI, essentially because it is the same technology reverse-engineered. For many doom-mongers, AI is the magic new ingredient because it suggests not so much an artificial intelligence as an independent one that escapes this constraint: "Because the algorithms that generate the fakes continuously learn how to more effectively replicate the appearance of reality, deep fakes cannot easily be detected by other algorithms — indeed, in the case of generative adversarial networks, the algorithm works by getting really good at fooling itself". This mistakes the brute force trial-and-error of a GAN for the subtlety of a con-man. The personalisation ("itself") is doing a lot of work here.

The US is at the forefront of AI research but this is seen as problematic because it is driven by the private sector (the traditional conservative suspicion of the market is still to be found in the arena of national security). The fear is that an irresponsible search for profit will allow foreign powers to exploit commercially-available AI tools to spread disinformation, just as it allowed them to leverage the automated advertising systems of social media platforms in recent years. What the state wants is greater control. As the Republican chair of the Senate Intelligence Committee said "one of the most important things we can do is increase information sharing between the social media companies who can identify disinformation campaigns and the third-party experts who can analyze them." Given that those third-party experts will be government-funded, this looks like a push to institutionalise a nexus between technology platforms and the state. That's not unusual - the same nexus has long existed between the state and the media - but it's worth noting that new media doesn't temper this cosy relationship with an invigilatory role over government in the way that old media does. The "platform not publisher" argument is presented as irresponsible in the current struggle, but what it really boils down to is an unwillingness to antagonise the government, indeed any government. As old and new media inevitably converge, that conservative tenor will become dominant.

With the global war on terror now winding down, there is obviously a need for a new enemy to justify the national security state. John Naughton, who as an Oxford academic contributes to the research on the Internet and conspiracy theories, characterised the implications of the two US Senate reports as "endless infowar", though he seemed to be thinking primarily in terms of the persistence of the Russian threat. He also quoted Kevin Roose of the New York Times who revealingly talked of social media in language that echoed the neoconservative framing of Afghanistan and Iraq in the early years of this century: "It’s the terrain on which our entire political culture rests, whose peaks and valleys shape our everyday discourse, and whose possibilities for exploitation are nearly endless. And until we either secure that ground or replace it entirely, we should expect many more attacks, each one in a slightly different form". Of course, that sort of language was not restricted to neoconservatives. Plenty of liberals drank the Kool-Aid and whooped and hollered from the sidelines. What's different today is that liberals are leading the campaign while the neocons (many of whom remain "never-Trumps") politely applaud.

Russia cannot credibly play more than a walk-on part in this drama. Inflating its impact by talk of asymmetric warfare and the strength of the weak is unconvincing. Outside of political groups looking for excuses, and newspapers with a vested interest in the restraint of social media, few people think that Russia swung either the US Presidential election or the UK's EU referendum. This is not to say that both weren't subject to disinformation and outright lying, but the more impactful sources were probably domestic, notably partisan TV and newspapers that had been pursuing a systematic campaign of disinformation for decades. A feature of the case against social media is the belief that it is much more effective than traditional media in influencing voters, even if there is scant evidence for this and the claim requires careful framing. For example, the New York Times conclusion that "In an election decided by a rounding error — fewer than 80,000 voters spread over three states — Russian trolling easily could have made the difference", becomes "Yes, Russian Trolls Helped Elect Trump" in the headline. This is no more justifiable than that infamous 1992 headline, "It's the Sun wot won it".

Given that Russia hasn't let up on its disinformation campaigns in the US, you would have expected the 2018 mid-term elections to have further entrenched the Republicans, but the reverse happened. Indeed, where the Democrats fell short, old fashioned voter suppression by administrative means was the more obvious explanation. Likewise, the investigations into "dirty money" and disinformation around Brexit have not notably influenced public opinion on the issue. Insofar as the polls have moved (slightly), this probably owes more to fears of a damaging no-deal outcome. Naughton is probably right that we face the prospect of a seemingly endless infowar, but the reason for that has less to do with Russian malevolence than the congruence of interest between an increasingly conservative media establishment and an increasingly liberal national security establishment. The former's defence of a supposedly beleaguered liberal order has seen it adopt not just reactionary manners - its obsession with civility and norms - but the paranoid style of anticommunism. The latter's shift away from the raw imperialism of the neocon years has seen it embrace the idea of a market-led projection of soft power, which is really just a revival of Cold War thinking. The response to social media has been nothing if not nostalgic.