Unethical & Manipulative Algorithms of Social Media Platforms

The algorithm used in social media is much more dangerous than AI/ Generative AI contents as this algo feeds the information to Generative AI. Everybody is speaking to have a governance or regulation around AI, no one thinks about the source of the problem within social media space.

The lack of regulatory guidelines for social media algorithms within platforms like Facebook, TikTok, X (Twitter), Instagram, Snapchat, Telegram - particularly those that drive users into coercive and directionless behaviour, poses significant ethical and societal concerns.

Below are key gaps and considerations that should inform regulatory development:

1. Transparency of Algorithmic Processes Social media platforms often lack transparency about how algorithms prioritize and recommend content. This opacity leaves users unaware of how their engagement is being influenced.

2. Ethical Use of Data Algorithms frequently use user data for predictive and persuasive purposes, often pushing harmful or addictive content without explicit consent or awareness.

3. Content Moderation Ambiguity Current moderation guidelines are inconsistent, and algorithms may amplify divisive, sensational, or harmful content for engagement purposes.

4. Content Amplification for Engagement Algorithms prioritize engagement over ethics, promoting sensationalist or explicit content because it captures user attention and extends screen time.

5. Lack of Accountability Mechanisms Platforms are not held sufficiently accountable for negative societal impacts such as misinformation, polarization, and mental health deterioration caused by algorithmic decisions.

6. Manipulation of Vulnerable Users Young users, individuals with mental health challenges, and those susceptible to addiction are often disproportionately targeted by algorithms designed to maximize screen time.

7. Bias in Algorithmic Decision-Making Algorithms may perpetuate or exacerbate biases (racial, gender, political, etc.) due to lack of robust oversight and diverse training data.

8. Lack/ Absence of User Control Users lack adequate tools to customize or opt out of algorithm-driven feeds, leaving them vulnerable to unintended manipulation.

9. Promotion of Directionless Content Platforms prioritize engagement metrics, leading to the promotion of trivial or non-constructive content that hinders intellectual and social growth.

10. Psychological Manipulation The algorithms use behavioral data to predict and exploit vulnerabilities, leading users toward addictive content, including sexually explicit material, regardless of their original intent.

11. Normalization of Harmful Behavior Prolonged exposure to pornographic or exploitative material can normalize unhealthy attitudes, distort relationships, and negatively impact mental health.

Proposed Regulatory Solutions

1. Mandate Transparency Reports Require platforms to disclose detailed information about how algorithms operate, including content ranking criteria and data usage policies.

2. Algorithm Audits Establish independent bodies to audit social media algorithms for fairness, ethical compliance, and societal impact. Require regular third-party audits to ensure algorithms do not promote explicit or harmful content disproportionately or unintentionally. Enforce stricter safeguards for minors, including age verification systems and prohibition of explicit content recommendation in their feeds.

3. User Empowerment Implement features allowing users to toggle algorithmic recommendations, view content chronologically, or customize algorithm preferences.

4. User-Controlled Content Filters Provide tools allowing users to filter out content categories they wish to avoid, including sexually explicit or psychologically triggering material.

5. Global Ethical AI Guidelines Introduce global standards for ethical algorithm design, focusing on minimizing harm and fostering constructive user engagement. Collaborate with international bodies to establish universal guidelines on ethical algorithm design, focusing on harm reduction and user protection.

6. Data Usage Consent Ensure explicit user consent for data collection and usage in algorithmic models, with clear opt-out options.

7. Stricter Content Moderation Standards Develop uniform guidelines to prevent the amplification of harmful, false, or divisive content.

8. Age-Appropriate Algorithm Regulation Design separates algorithmic frameworks for minors, emphasizing educational and developmental content over addictive engagement.

9. Psychological Impact Assessments Require platforms to conduct studies on the psychological effects of their algorithms and make these findings public.

10. Penalty Framework for Non-Compliance Impose fines or sanctions on platforms that fail to control algorithmic promotion of pornographic or manipulative content.

11. Explicit Content Detection and Demotion Use advanced AI models to identify and demote pornographic content systematically, ensuring it does not reach unintended audiences.

Case Study: Potential Models for Inspiration

1. European Union's Digital Services Act (DSA) The DSA imposes strict accountability on platforms for content moderation, transparency, and user safety, including specific rules to prevent harmful algorithmic effects.

2. UK's Online Safety Bill This legislation emphasizes protections against harmful content, particularly for children, requiring platforms to mitigate exposure to explicit material.

3. Australia’s eSafety Commissioner Proactively targets online harm by working directly with platforms to address risks, including the promotion of explicit content through algorithms.

The Role of Governments and Industry Stakeholders

• Governments should collaborate with industry experts, academics, and civil society organizations to draft robust guidelines.

• Platforms must be incentivized to adopt ethical practices voluntarily, with regulatory oversight ensuring compliance.