A world where AI and humans live in symbiosis. Where we are part machine and part human, with every quality a human can have but more advanced in functions. A future where none of us dies but lives forever, continuously evolving. This could be the potential, more optimistic view on AI and homo sapiens.
Steven Hawking once said that “the development of full artificial intelligence could spell the end of the human race.”
The Terminator-like situation that Steven Hawking depicted is a warning in itself. Here, if we do not evolve with AI, AI could potentially overthrow us and begin an extermination phase.
It was once theorised that the Neanderthals disappeared (leaving behind some lineage, though) due to mass annihilation by Homo Sapiens. As we have seen throughout history, when a culture is deemed a minority, the bullies come and segregate or, in the worst-case scenario, begin a mass extinction (much like what happened to Indigenous populations).
AI could go down three pathways. One being the aiding of humans and being treated as an assistant (hello Alexa). The second being mass extinction of humankind (“I’ll be back”). Additionally, the third path is the one where humans and AI achieve symbiosis.
AI is a wonderful field, but little is known about it beyond generating content from prompts. This is the topic of every news article, and the noise hasn’t died down. Either the AI models could now detect a malignant tumour that would otherwise go unnoticed before the stage 4 diagnosis, or the invention of a supercomputer with quantum computing added to the mix. The possibilities are endless.
AI is a very broad field. AI consists of different subfields such as machine learning and deep learning. I am currently learning Machine Learning, and boy has it been a massive rabbit hole. During the first five weeks of the unit, I have learnt how to do data wrangling and have yet to do linear regression and other actual machine learning coding.
Given this oversight into the breakdown and narrow components of AI, it can illustrate the complexity of AI. The closer you get to the centre, the deeper into the rabbit hole you go (hopefully you get to have tea with the mad-hatter).
Most AI models currently function in a complex black-box setting, which makes it harder to understand and reconfigure. Additionally, there are biases and discrimination arising from training data due to unnecessary pattern creation leading to prejudices. For example, the US criminal court system implemented an AI model to calculate the re-offending rate, but it resulted in the model being biased against African-Americans. Another example is Amazon abandoning the experimental AI model that was trained over ten years in a male-dominated field, which depicted bias against female candidates. These types of examples are the reason for human intervention and oversight with proper frameworks (such as DRIFT) to maintain and create AI models.
However, prejudices like these suggest that AI is susceptible to certain biases and discrimination that, if not maintained and minimised, will turn into the next Hitler (much like Microsoft’s teenage girl-based AI agent, Tay, who became very racist and biased within 16 hours of launch).
So, it is yet to be determined which (barely) futuristic path AI models are headed down. The Theory of Everything could determine The End of Everything. Where we find answers to all our questions but fail to save ourselves in the end…
Leave a comment