I've been traveling recently, which for me is a delightful mix of discovery and a stark reminder of how utterly dependent I am on the digital world. This isn't about the questionable joy of doom- scrolling world events from a café in a picturesque town, but about the fundamental mechanics of modern travel. On a train journey through Germany, an email pinged to inform me that my connecting train had been cancelled. Armed with this digital heads-up, I was able to find alternatives online and even help a fellow couple who were blissfully unaware of the impending travel chaos.
This reliance became a recurring theme. My bank, smelling the unfamiliar scent of international transactions, decided to block my card for suspected fraud (it wasn’t), and unblocking it required a digital pilgrimage through online portals and authentication apps. A hotel sent our door PIN via WhatsApp. My wife had a minor medical query that a quick online search thankfully confirmed didn't require a doctor. Each incident reinforced a now almost universally accepted truth: the ability to get online to send an email or look something up is no longer a luxury, but a basic necessity for navigating modern life. Some might even call it a minimum human right.
During these same travels, I found myself leaning on AI quite heavily. It became my on-the-fly translator, my personal botanist, ornithologist, and even my guide to the intricate workings of cathedral organs. I used it to understand the historical confluence of the Danube and Inn rivers, learning how the decline of the salt trade led to the planting of mulberry trees in Passau.
Professionally, I even used it to understand why my dreams were so vivid (it turns out high altitude can have that effect). None of this was world-shattering, but it was incredibly useful and, dare I say, fun.
I know I'm an outlier, a heavy user of AI in my research and daily life. But looking ahead, it’s not hard to imagine a future where most of us will need AI to translate a complex document, whether from technical jargon into plain English or from one language to another.
This leads me to a critical question, particularly through the lens of my work in digital health equity: Is it time to discuss a minimum set of AI capabilities as a basic right for everyone, much like we now view internet access?
The Case for a Minimum AI Toolkit If we were to define a pragmatic "minimum set" of AI capabilities that everyone should have access to, what might it include? Here are a few thoughts to start the debate:
- Translation and Comprehension: The right to understandable information is crucial, especially in public services, healthcare, and emergencies. This would mean access to text and speech translation and tools that can simplify complex official documents. The United Nations' Universal Declaration of Human Rights, in Article 27, states that "Everyone has the right... to share in scientific advancement and its benefits".
- Accessibility Support: AI can be a powerful tool for people with disabilities, offering captioning, image descriptions, and summarization. This aligns with guidance from the Office of the United Nations High Commissioner for Human Rights on AI and the rights of persons with disabilities.
- Civic and Safety Uses: In a world of disruptions, having AI that can help parse official notices, disaster alerts, and travel updates is invaluable. The recent UN General Assembly resolution on AI emphasizes the need for safe, secure, and trustworthy AI that is accessible to all, particularly in developing nations.
- Privacy and Dignity Safeguards: Any right to access AI must be coupled with the right to opt-out of biometric categorization and protection from manipulative systems. The EU AI Act is a significant step in this direction, establishing risk-based duties to protect fundamental rights.
- Continuity of Access: This right would be meaningless without protection from arbitrary denial through internet shutdowns or blanket blocks, a point repeatedly emphasized in UN Human Rights Council resolutions.
The Case Against Declaring a New Right
Now, for the other side of the coin. The argument against creating a new "AI right" is compelling and deserves serious consideration.
- Dilution of Core Human Rights: Critics argue that we don't need a new right. Existing frameworks already protect access to information, non-discrimination, and health. The focus, they say, should be on rigorously applying these existing rights to the age of AI, not creating a new category that could dilute the meaning of human rights.
- Risk of Deepening Inequality: If not implemented carefully, universal access to AI could actually worsen societal gaps. Civil society organizations like the ACLU have warned that biased models and the potential for surveillance can exacerbate disadvantages, especially if robust privacy and fairness protections are not in place.
- Governance Gaps and Misuse: There is a real concern that a right to AI could be co- opted for manipulation or repression, a risk documented in analyses for the European Parliament. Without strict limits on abusive systems, this could be a double-edged sword.
- Fragile Infrastructure: Many "AI rights" would depend on internet connectivity, yet internet shutdowns remain a common tool of control for many governments, as tracked by organizations like Access Now. However, we are seeing a rapid rise in powerful, small AI models that can run offline on a laptop or smartphone, which could mitigate this concern.
- Copyright and Ethical Concerns: The legal and ethical landscape of AI is still the Wild West. Ongoing battles over training data, combined with the immense energy consumption of training large models, raise significant questions about the sustainability and fairness of the entire ecosystem.
Where Do We Go From Here?
The path forward is anything but clear. If we leave it to the market, the outcome is predictable: the knowledge-rich will get richer, and the knowledge-constrained will be left further behind.
The UNESCO Recommendation on the Ethics of AI provides a framework, emphasizing human oversight, transparency, and fairness.
So, where do you stand? Is it time to advocate for a universal set of AI capabilities as a fundamental right, or should we focus on strengthening our existing human rights frameworks to meet the challenges of the AI era?
What’s the one AI capability you believe everyone should have access to, and why?
Sources
Here's some sources:
- CESCR General Comment No. 25 (2020): Authoritative guidance on states’ obligations to ensure equitable access to the benefits of scientific progress, including targeted measures for groups facing systemic exclusion. OHCHR+1
- UN GA AI resolution (21/03/24): Calls for safe, secure, trustworthy AI with equal access, especially for developing nations. AP News
- EU AI Act: Risk-based duties to protect fundamental rights; bans and obligations for high-risk use. EUR-Lex
- Australia’s universal-service settings: Movement toward a single broadband and voice baseline for all premises; ACCC submission on ensuring availability, accessibility, and affordability. Infrastructure Australia+1
- Cautions: ACLU on inequality risks; Access Now trend of shutdowns; EP study on AI as a tool of manipulation. American Civil Liberties Union+2The Guardian+2
Leave a Comment