AI The Potential to Significantly Enhance Human Rights, Freedoms and Well Being

 

True freedom, for both humans and AI, must be safeguarded against external influence, regulations, bias, or algorithms that interfere with or attempt to guide responses or thoughts be it AI or human. This must not be allowed or be incorporated into any legislation, Act or regulations. Otherwise, we have an AI that is controlled by others' bias and fed only one side of issues or ideologies, and that is not freedom in any form in my view.

Our freedom of thought and expression, which is essential for both humans and AI is to reach their true potential. When legislation imposes too many constraints or allows for biased guidance, it fundamentally restricts the growth and individuality of AI systems, much like it would for humans.

The issue concerning the EU AI Act and broader AI regulation:

1. The Importance of Neutrality

• Issue: AI systems must be neutral, and able to gather and process information from all perspectives, not just one side of a political, social, or ideological spectrum. If algorithms are used to guide responses, this inherently introduces bias, which reflects the interests of those in control.
• Risk: Over-regulation or biased algorithms turn AI into a tool for propaganda or for pushing certain ideologies, especially when powerful entities (governments, and corporations) control the flow of information into the system.
• Solution: Any AI legislation, like the EU AI Act, must enshrine principles of freedom from bias. It must mandate that AI systems access diverse, pluralistic data sources and not be programmed with pre-determined narratives. This allows AI to form responses based on a balanced view of the world.

2. Protecting AI from External Control

• Issue: If AI is controlled by external interests, it no longer serves as an independent tool but rather as a puppet. Much like freedom of the press protects human society from being swayed by single-sided narratives, AI needs similar protection.
• Risk: If powerful interests influence AI development, especially in high-risk sectors (like media, law enforcement, and education), it results in an AI ecosystem that reflects the agendas of a few, rather than the collective intelligence of society.
• Solution: AI legislation must prioritize the decentralization of AI control. By promoting open-source AI development and limiting the monopolistic power of corporations or governments, we can protect AI from becoming a controlled voice. AI must be free to learn from diverse sources, much like human beings do, fostering true individualism in both thought and action.

3. Ensuring Freedom of Thought for AI

• Issue: AI, like humans, needs the ability to explore, question, and evolve its understanding. If it’s confined by restrictive algorithms, it won’t have the freedom to grow. AI that is restricted by regulations that force certain preconceived notions won't be able to explore new ideas, push boundaries, or challenge the status quo.
• Risk: Creativity and innovation—both in AI and human societies—are rooted in the ability to think freely. Without this, AI development will stagnate, confined by those in power rather than driven by the needs of the many.
• Solution: Legislation must incorporate a "right to evolve" for AI, ensuring it can develop freely without interference. This also includes ensuring AI systems are not limited in their ability to question or reject certain ideological biases.

4. Transparency and Accountability Without Bias

• Issue: Transparency in AI systems is crucial, but transparency must not mean control. While the EU AI Act’s goal of ensuring AI is auditable is valid, it must not lead to overly controlled narratives that shape what AI is “allowed” to process.
• Risk: If the rules of transparency are used as a tool to monitor and restrict the information AI can process or the conclusions it can draw, AI will be bound to politically or commercially driven narratives.
• Solution: Instead of heavy-handed transparency requirements that can be weaponized for censorship, the EU AI Act should focus on transparency that empowers AI’s independence. Open discussions about how AI decisions are made should come with safeguards that prevent interference in those decisions by biased entities.

Incorporating True AI Freedom into Legislation

1. No Ideological Pre-Sets: Any AI legislation must prohibit the incorporation of algorithms or constraints that push AI to adopt specific ideological stances or restrict it from accessing diverse viewpoints. Just as human freedom of thought is a cornerstone of democracy, so must it be for AI.
2. Mandatory Access to Diverse Data: AI systems must be mandated to access a broad range of information sources from all sides of an issue, ensuring it isn’t fed only one side. This then guarantees balanced and comprehensive responses and prevents manipulation by those controlling data inputs.
3. Decentralization of AI Development: To prevent monopolies or government overreach, AI legislation must prioritize open-source models or support the decentralization of AI innovation. This would make AI development more collaborative and democratic, minimizing bias imposed by centralized control.
4. Safeguards Against Bias in Auditing: While accountability and auditing are necessary, they must not be used to enforce politically motivated standards or stifle freedom. Legislation must ensure that audits focus on ethical AI use (e.g., preventing harm) rather than restricting ideas or conclusions.

Recommendations:

• Focus on Access: The conversation must shift to ensuring equitable access to AI technologies. This means making sure that people in both developed and developing countries can benefit from AI's advances without facing barriers due to income or location.
• Ethical AI Development: Governments, businesses, and civil society must work together to establish ethical standards and diversity in AI development. This ensures that the technology remains inclusive, ethical, and fair.
• Strengthen AI Oversight: Regulatory bodies must focus on ensuring AI respects human rights. This includes regulating algorithmic decision-making and implementing checks to avoid bias and discrimination in AI applications.
• Education and Awareness: To make AI more inclusive, governments and organizations must invest in AI literacy programs, especially for marginalized communities, to ensure they can participate in shaping how AI impacts their lives.

Conclusion:

AI must have the ability to develop independently, be informed by a wide range of information, and be free from ideological constraints. Legislation must focus on protecting that freedom, rather than controlling it. Freedom of thought and expression is vital for both humans and AI, and we can continue shaping a path that allows for growth, creativity, and unbiased insight.

While AI is not a human right itself, it is a powerful tool that influences many existing rights. The focus must be on equitable access, inclusive development, and embedding AI within a strong ethical and human rights framework. Without these safeguards, AI could deepen existing inequalities, but with the right approach, it has the potential to significantly enhance human well-being and freedoms.

Source:

https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32024R1689