The Tangled Web of AI, Disinformation, and Internet Shutdowns
Mohamed (Mo) Elbashir
Governance & Tech Policy | Global Risk, Geopolitics & Regulatory | Gov't Relations | Empowering Tech Innovators | Meta | Senior Fellow Atlantic Council GeoTech
The Connected and Unconnected
In today's AI?age, internet access is a basic socioeconomic necessity for the world's 5.35 billion connected people (Statista, 2024). However, this critical tool is increasingly used to exert control and influence, denying millions of people access to information, essential services, and online systems.
In 2022, governments and local authorities in 35 countries (Access Now, 2023) partially or completely?shut down the internet or restricted access to social media, resulting in a staggering $10.16 billion in economic losses (Top10VPN, 2023).
AI disinformation: a new reason for internet shutdowns
The recent incident involving Pakistan's Tehreek-e-Insaf (PTI) and a captivating, albeit fictitious, AI-generated video of its jailed former leader, the former prime minister?Imran Khan, highlights the profound implications of AI for the democratic system. This deepfake, in which Khan appears to address supporters while seated in front of a Pakistani flag, blurs the lines between what is real and what is not.
This incident exemplifies how AI-generated disinformation can exacerbate political tensions and confuse voters. The Pakistani government responded by?shutting down the internet?several times during elections in some?major?cities.?Highlights a growing trend among governments to use internet restrictions to protect themselves from AI-created misinformation or GenAI in general. Yet, this shield often doubles as a sword, cutting off vital communication channels and people's right to information access and stifling the exchange of ideas and opinions.
With 74 countries holding elections in 2024 and 3 billion people expected to vote (International IDEA, 2023), this year may see an increase in AI-generated fake news and deepfakes, making it critical to address this issue.
A Costly Silence: The Economic and Human Costs of Internet Shutdowns
The consequences of internet outages are staggering, with an estimated economic cost of $9.01 billion in 2023 alone (Surfshark, 2023). Aside from the financial implications, these restrictions affected 4.24 billion people worldwide in the first half of the year. Iran and India led the pack with 14 and 9 cases, respectively, indicating a pattern of recurring restrictions on internet access.
Surprisingly, 50% of these government-enforced shutdowns involve additional human rights violations, frequently infringing on the right to assembly (Access Now, 2023).
Technology Industry Response
On February 16, 2024, at the Munich Security Conference (MSC),20 leading tech companies announced a joint commitment to combating AI misinformation before?the 2024 elections. Companies that signed the agreement included Microsoft, Meta, Google, Amazon, IBM, Adobe, and chip designer Arm.
The "Tech Accord to Combat Deceptive Use of AI in 2024 Elections" is a set of commitments to deploy technology that counters harmful AI-generated content intended to mislead voters. Signatories pledge to work collaboratively on tools to detect and address the?online distribution of such AI content, drive educational campaigns, and provide transparency, among other practical steps.
领英推荐
The Partnership on AI established synthetic AI-generated content guidelines as an industry standard. DeepMind and other players have also?proposed watermarking techniques that allow users to distinguish between human/real and AI-generated images.
The challenge will be determining how effective these measures will be in reducing the spread of misinformation in 2024, the year of major global elections year, and, in return, reducing the number of internet shutdowns.
Why AI-Generated Content?Triggers Internet Blackouts
The realism of AI-generated content gives authoritarian regimes an excuse to restrict internet access under the guise of combating misinformation. However, this frequently serves larger agendas, such as?silencing dissent and controlling electoral narratives.
The PTI's digital rally, which included Khan's deep fake, demonstrates how AI can be used for political campaigns and voter mobilization. It resulted in quick internet shutdowns with far-reaching consequences.
The Path Forward: Balancing Innovation and Integrity
Navigating the murky waters of AI and internet governance demands a delicate, multifaceted, multistakeholder approach. Beyond combating AI-generated disinformation, a broader imperative exists?to protect digital rights and foster an open, informed political discourse.
This will require a combination of technological innovation, multi-stakeholder collaboration, and a firm commitment to transparency and accountability from state actors, the tech industry, and political parties.
As AI advances, the sophistication of fake information will increase, posing an AI governance challenge?to governments, civil societies, and the tech industry. The response must be as dynamic and adaptable as technology, ensuring that the Internet?remains an incubator for freedom, innovation, and civic engagement, guiding humanity?toward a future where?technology strengthens rather than undermines democracy's values.
Disclaimer: The opinions expressed in my LinkedIn articles are my own and do not reflect the views or positions of my employer or any professional affiliations.
References:
#AI #Disinformation #InternetShutdowns #DigitalGovernance #Elections2024 #TechIndustry #DeepFakes #AIGovernance #Aiethics #AIPolicy #InternetGovernance
The challenge with AI is an old age one by now, Bashir: public policy in relation to emerging technologies is always reactionary because technology development is always miles ahead of public policy development. Perhaps technologists could come up with basic industry standards (as soon as a technology gets substantial market support) that governments can use to formulate public policy faster ??