AI and Liberty, A Topic Less Tantalizing

Kahlil Gibran is credited with saying, "Life without liberty is like a body without spirit."

My question is, what will human existence be in the age of brilliantly performing human-like machines?

I spoke as a panellist on AI and Liberty at a recent event organised by the Mont Pelerin Society in New Delhi on September 23. I had the distinct impression that this topic is perceived as far less exciting than AI in Education, health, etc. Understandably, we humans get easily caught up in the fantastic potential of miracle machines. We would rather marvel at the speed with which technologists are imbuing machines with human-like capacities than dwell on the downside or see troublesome patterns.

I felt like an event/party pooper when I spoke about the similarities between our approach to progressively losing our privacy online, being sucked into a brave new world of social media-guided self-worth, notions and relationships and the simultaneous loss of human agency and attribution of human rights and attributes to machines.

This is not a binary issue. I believe it is perfectly possible to reap the full potential of AI as a technology while avoiding past mistakes and delays in tackling harm arising from excessive personal data collection, behavioural manipulation, and exploitation. We need a different approach and business models. The latter is the difficult part.

I touched upon the various myths and language surrounding AI that we have willy-nilly accepted as a given. For example, we have become accustomed to saying that AI thinks, creates, and learns. I firmly believe that, for starters, there should be a ban on using the grammar that describes human capabilities and human actions to describe what machines can do.

For example

·      No, AI does not "hallucinate". It gives defective results because of poor data inputs, design and incomplete training.

·      AI does not "learn" from our papers and content. Unless it is made to acknowledge and reference, that is plagiarism by the tech company that created it.

·      AI is not educated. It does not "learn or self-learn"; it is human-programmed software. We should say AI is programming itself if it can.

·      AI is not conscious, so if it acts like it is, it performs as programmed.

Being free from false narratives will be crucial for the future of humans in the age of AI.

I concluded by stating that I would like the discourse to shift to how we can redirect technology towards shared values and new business models that are not based on data theft and manipulating humans and do not sell us myths that diminish our humanity's agency and liberty.

This calls for a grand re-education of all stakeholders and a demand for a new set of values to underpin human existence.