Shaping intentions, misinformation, and filter bubbles: AI's double-edged sword ⚔️

Is AI shaping our choices or saving us from ourselves? The future of AI isn’t just about its power, it’s about how we choose to use it!

Jan 10, 2025

Is AI shaping our choices or saving us from ourselves?

Imagine a world where AI doesn’t just observe your choices, but actively shapes them! From crafting personalised content to generating realistic deepfakes, AI has revolutionised how we consume information… but it also poses serious risks.

Unchecked, AI can manipulate decisions, spread disinformation, and trap us in echo chambers, eroding trust and democracy. Yet, it also holds the potential to fight back ⇒ detecting falsehoods, promoting diverse perspectives, and fostering digital safety.

This article explores AI’s double-edged role in disinformation, filter bubbles, and the rise of the intention economy. The future of AI isn’t just about its power, it’s about how we choose to use it!

💡 Disinformation costs

The global economic cost of disinformation exceeds $78 billion annually, including impacts on elections and public health.

Disinformation at scale

AI has supercharged disinformation, enabling the rapid creation and spread of false narratives through large-scale campaigns that manipulate public opinion with unprecedented speed and precision. These tactics, which include swaying elections with fake account on social media (that can now seems very real with their own personalities etc.), targeting marginalised groups, and creating hyperrealistic deepfakes, erode trust in institutions and deepen societal divides.

Deepfakes (AI-generated fake audiovisual content) are a particularly potent tool in this arsenal. From forging documents to discrediting public figures, their growing sophistication poses significant challenges. The emergence of “Deepfake as a Service” (DaaS) has made these tools widely accessible, further amplifying their impact.

Filter bubbles and polarization

On social networks and media, AI is used everywhere to recommend new posts, videos, articles, etc. This AI personalised content often traps users in filter bubbles (self-reinforcing environments that limit exposure to diverse views). Those echo chambers amplify biases, deepen divides, and, in extreme cases, foster radicalisation.

The result? Polarization that undermines democratic dialogue and informed decision-making! This is, in part, what we see with the polarization of politics in Europe and the United States in particular.

Furthermore, large language models like ChatGPT or Claude can unintentionally reinforce this issue, tailoring responses to align with perceived biases. Users may not realise their information is filtered, further narrowing perspectives.

Misinformation flourishes in filter bubbles

Misinformation spread more easily in filter bubbles, where users are more likely to accept false or inaccurate information that aligns with their existing beliefs. This creates a feedback loop: misinformation and misinformation reinforces biases, and filter bubbles amplify its impact, further narrowing perspectives and deepening divides.

Breaking this cycle requires transparency, moderation, and media literacy. By addressing these interconnected issues, we can reduce the harm of AI-driven content systems and promote informed online engagement. But as AI evolves, so do its methods of influence, leading to the rise of the intention economy.

The rise of the intention economy

The “attention economy” — where companies compete to capture and hold people’s focus, often through engaging or addictive online content, to drive advertising revenue — is evolving into something more powerful: the “intention economy,” where AI predicts and shapes your choices.

Advanced tools, like large language models (LLMs), don’t just suggest content ⇒ they guide decisions, from purchases to political preferences, based on your behaviour and mood. This power isn’t just theoretical. Meta’s AI Cicero predicts intent in strategy games, and Nvidia envisions AI anticipating users’ every move.

While promising convenience, these technologies risk covert manipulation, allowing corporations and political entities to steer behaviour without consent. Unchecked, the intention economy could deepen inequalities, undermine personal autonomy, and compromise democracy.

AI is shaping our views and then our choices

To summarise what we have seen, AI is enabling disinformation on a massive scale, creating social bubbles that narrow our view of the world and radicalise our opinions. In these social bubbles, fake news and misinformation flourish more easily, creating a reinforcing feedback loop. So our views and experiences are being shaped by AI, which is already indirectly shaping our behaviour. But it goes further, as AI then uses all this knowledge of our opinions (which it has helped to shape) to anticipate our decisions and directly shape our behaviour.

On this bad note, how can we ensure AI serves as a force for good rather than manipulation?

“Fightback AI” as a solution

While AI contributes to disinformation, misinformation, filter bubbles and behaviours manipulation, it also offers powerful tools to counter these issues. Advanced systems can detect falsehoods by analysing patterns and context, flagging harmful content before it spreads. However, deepfakes technics evolve quickly, making detection an ongoing race between creators and defenders.

Solutions include improved authentication processes, such as live video verification with randomised actions. Initiatives like the Coalition for Content Provenance and Authenticity (C2PA) use watermarking to verify content authenticity, providing a robust defence against manipulated media. Tools like Nobias, Grounds News, or non-profit like USAFacts, help users track bias in news consumption and diversify their information sources, encouraging balanced views.

Building a future of trust

AI’s dual role as both a risk and a remedy underscores the importance of deliberate choices in its development and use. However, technology alone isn’t enough. Public education in critical thinking, media literacy, and collaboration among policymakers, law enforcement, media, tech companies, and educators are essential. Together, these efforts can rebuild trust and foster a digital ecosystem where truth and diversity prevail.

Regulation plays a critical role, as seen in the EU’s risk-based AI framework. This approach mandates labelling manipulated content and classifies deepfakes detection tools as “high risk,” requiring rigorous oversight. Europol’s Innovation Lab is also equipping law enforcement with the skills and tools needed to detect and counter deepfake-related crimes.

The challenges of disinformation, filter bubbles, and manipulation are significant, but so are the opportunities to counter them with innovation and cooperation. By harnessing AI responsibly, we can ensure it uplifts society, safeguards democratic values, and fosters an informed, inclusive digital world. The future of trust in technology depends not on the tools themselves, but on how we choose to use them today!

⚠️ Practical tips to stop disinformation

Watch for inconsistencies in style, tone, or visuals.
Use AI tools like reverse image searches to verify authenticity.

Artificial Impact

Discussion about this post