Futurist Keynote Speaker: Posts, Slides, Videos -
Artificial Intelligence / Robotics / Cyber etc
What will be the future impact of AI / Artificial Intelligence on our wider world? The answer is that it will be massive in every dimension of human life.
Most of the influence of AI on your own life will be hidden, and hard to understand but the total combined effect will be very powerful, affecting decisions you make, influencing your thoughts, shaping your opinions, affecting your own unconscious bias and preferences.
To help explore these issues, I thought it would be revealing to have a conversation (below) about the future of AI with AI itself - using ChatGTP, a powerful Artificial Intelligence engine, which has learned about the world by analysing billions of words written by human beings. My conclusion about the conversation is that ChatGTP is not telling the truth when it comes to the downsides of AI.
Word-based machines like ChatGPT are running out of text to learn from
Word-based AI machines like ChatGPT are already up against a limit in their training programmes. For example, ChatGTP has already read most of the available words online and is rapidly running out of new material to learn from! So the next step-wise improvement will need to come from other innovations in Artificial Intelligence, and from actually engaging with live human beings.
We can expect another major extension of AI language learning by feeding AI with recorded content - all the world's TV, radio and films for example. But the total number of new words in all that is surprisingly small - human beings can't speak much faster than 130 words a minute, so even an hour of TV rarely contains more than 6000 words.
Bear in mind that I was talking only to Version 1 of ChatGTP, that ChatGTP has no knowledge of the world beyond 2021, and that the engine is set to deliver short answers only, the results are very impressive.
So what of the future? How good will such word-based engines be by 2030? Answer: good enough to convince most people most of the time that they are chatting online to a human being.
AI is already good enough to pass some student exams
The AI system is already good enough to deliver coursework for exams, complete homework assignments, write press releases or information bulletins, write computer code, and so on.
Expect billions of paragraphs to be generated by AI bots like ChatGTP, potentially swamping the web and search engines with automated articles, social media posts and so on. This will have two effects.
Firstly it is likely to degrade the experience of human beings searching for world-class content, amongst all the automatically generated clutter.
Secondly, the text produced by AI will be swept up into the pool of all available text to train all AI chatbots - so that AI text begins to train AI to produce more AI text. This could easily result in a distortion, subtle at first, for example in literary styles. It will mean that early AI bots may have a profound impact on the language capabilities and "personality" of bots that follow.
So here below is the AI view of the future of AI. You will notice that many of the answers are quite upbeat and tend to avoid discussions of risks from AI, threats in a world dominated by Artificial Intelligence. Text created by ChatGTP is in italics. My own comments are in bold....
Read more: How AI / Artificial Intelligence will transform every industry and nation - banking, insurance, retail, manufacturing, travel and leisure, health care, marketing and so on - What AI thinks about the future of AI? Threats from AI? Keynote speaker