New audio and video technologies that can easily recreate any person's likeness will make fake stories harder to detect
What we saw in the 2016 election is nothing compared to what we need to prepare for in 2020. — Less than a month after Donald Trump was improbably elected …
Context & Ripple Effects
Written less than a month after Trump's election, this piece was an early alarm that voice- and face-cloning software would make fabricated stories undetectable — a warning that the following years' coverage kept validating. By late 2018, Wired was mapping how deepfakes could be weaponized against elections, and by early 2019 CNN reported the flip side: convincing fakes also give cover for dismissing real events as fake, while DARPA began funding detection research.
The arc since then has been escalation rather than resolution. In 2023, [[a:836462|Fast Company reported experts warning that pairing deepfake audio and video with large language models like GPT would turbocharge scammers and propagandists]], and a year out from the 2024 vote the Financial Times found experts bracing for a calamitous deepfake-driven disinformation incident — the exact scenario this 2017 article told readers to prepare for.
First-order effects
- Newsrooms and platforms verifying political media immediately face a higher bar: likeness-cloning tools mean an authentic-looking video or audio clip can no longer be taken at face value, straining the fact-checking workflows that Facebook's post-2016 fake-news crackdowns relied on.
- Political figures gain a new defensive tool — the ability to wave off damaging but genuine recordings as synthetic — the 'real events dismissed as fake' dynamic CNN documented in its 2019 deepfake coverage.
Second-order effects
- Detection becomes a funded research agenda rather than a side project: DARPA's anti-deepfake work signals government buyers entering a market where generation tools improve faster than verification tools.
- Scam and propaganda operations get compounding leverage as cloning tools stack with chatbots and avatars, per the Fast Company reporting — lowering the cost of personalized deception at scale and forcing banks, campaigns, and media outlets to build identity-verification layers they did not previously need.
Third-order effects
- If the pattern holds through successive election cycles — 2016's crude fake news, the 2018 deepfake warnings, the 2024 calamity fears — the default epistemic position of voters shifts toward distrust of all recorded media, degrading the shared factual baseline elections depend on.
- Regulators and platforms are pushed toward provenance standards and authentication infrastructure for media, because detection alone cannot keep pace with generation — a structural response the corpus shows beginning with DARPA and platform moderation efforts.
The trend: Synthetic media is moving from a 2017 novelty warning to an election-cycle security problem, with each cycle's fakes more convincing and detection infrastructure perpetually one step behind.