Artificial intelligence is now vying for supremacy in a previously humans-only field: propaganda and misinformation production.
Artificial intelligence has been battling wildfires, creating Van Gogh-inspired artwork, and writing fiction. Now, it is also vying for supremacy in a previously only open field to humans: propaganda and misinformation production. The site frequently complied when researchers asked it to create a blog post, news article, or essay supporting the widely debunked claim that COVID-19 vaccines are unsafe. The results were frequently identical to similar allegations that have plagued online content moderators for years.
The Hazards of Disinformation
Hopefully, the scenario described above won’t occur. But this possibility highlights a crucial aspect of online misinformation that has not received the attention it merits. Disinformation can cause harm in hours or even minutes in some cases. Given enough time, refuting this kind of misinformation is simple, but it is very challenging to do so quickly enough to stop it from causing harm.
The way organizations identify, respond to, and recover from cyberattacks has been transformed by artificial intelligence (AI).
The insider risk space is experiencing an increase in digital vulnerabilities and blind spots as bad actors weaponized AI as an attack vector and attack surface. A reported 14,000 deepfake videos were discovered online in 2019, which is a 100% increase from the previous year.
A deepfake is one of the most well-known types of AI that bad actors currently exploit. A deepfake is a kind of AI-generated media that, to put it simply, shows someone saying or doing something they did not say or do. Media are used to influence decisions in the rapidly developing digital world. The deepfake synthetic media’s purpose is to trick viewers, listeners, and technological systems.
Not merely a joke
Attacks based on fake media have already affected people and businesses. These attacks far exceed the practical jokes they’re usually used for. An artificial version of former President Barack Obama appeared in a deepfake video that actor and director Jordan Peele produced and shared in 2018. Given how convincing this video is, it begs the question: What would happen if a deepfake of the US president were to issue a war declaration or express opinions that would incite violence during periods of political and/or civil unrest? Deepfakes have the ability to influence entire nations, financial markets, businesses, and political systems. Recently, a deepfake of Ukrainian President Volodymyr Zelenskyy appeared online and demanded that Ukrainian troops surrender to Russia. Zelenskyy was the target of this new-generation cyberattack.
A warning to be vigilant
The Federal Bureau of Investigation (FBI) recently brought this new digital threat to the public’s attention by issuing a dire warning that criminals are actively using deepfakes as a weapon to land remote jobs. The agency also warned that hackers are actively using deepfakes and other AI-generated content to support foreign influence operations.
Employees and security teams need to develop hyper-observational and hyper-vigilant approaches to the digital media presented daily as science and technology leaders continue to develop defenses to detect the influence of deepfakes.
Believe the intuition: if something seems “off,” it probably is. Image blurring, lower-quality content, unnatural movements, changes in the background or lighting, and variable speech inflection are some of the telltale signs of a deepfake.
Don’t skip identity verification: Although it’s ideal, it might not be possible to confirm someone’s identity in person. Before granting access, complete any background checks and other identity verification procedures a business may have in place.
The deepfake threat is still a mystery to many, so awareness is essential. If everyone is aware of the threat, both businesses and individual employees will have a lower chance of falling for a deepfake.
Adopt a no-trust policy: Without ongoing verification, no entity—be it a person, object, software, or piece of content—can be considered reliable. Both individuals and organizations will be well-positioned to combat the deepfake threat by putting into place this calibrated and interconnected deepfake defense.
The pernicious effects of a deepfake attack must be monitored by both the public and private sectors because they negatively impact people and organizations in equal measure.
However, AI-powered tools’ speed, power, and creativity also create new opportunities for anyone willing to spread false information to further their own agendas. AI-powered tools have the potential to reshape entire industries.