AI (Artificial Intelligence) is undoubtedly the future of computing and automation. It’s now receiving its well deserved recognition these days. From small scale businesses to large enterprises, every organization wants to be early adopter of this ‘genie in a bottle‘ tech without any foresight about its long term effects.
The applications of AI are limitless, though only LLMs (large language models) which are branded as ChatGPT 4, Bing Chat, Google Bard, DeepMind are getting all the hype. Image generation/manipulation AI such as MidJourney, Dall-E 2, Fotor and their countless other alternatives are also popping up every now and then on GitHub.
How AI Works
Do you remember Tasmanian Devil (Taz) character from Looney Tunes cartoon? He was hungry with never ending appetite and could eat anything that comes his way while spinning and grunting. AI works in a similar way, it feasts on information through through carefully programmed tools by developers using methods such as Neural Networks, Deep Learning and so on. Nowadays, AIs are feeding (training) on every data, text article, image, every photo that you upload to the internet, it doesn’t matter whether you grant them permission or not. They do not honor consent, much like Taz monster.
After building a collection (dataset) of publicly copied/scraped data from websites and applications, the data gets fine tuned using different models and filters are applied which prevent it to respond to certain inappropriate, illegal prompts by users. When a user types something as a prompt on ChatGPT4, it sequentially calculates relationship between the entered words using super computers (having quadrillions of transistors) and then answers (predicts) accordingly.
Currently, AI is roaming in a lawless territory, there are no ethics/legislations done so far. Governments around the world are unsure how to proceed with its regulation, what kind of threat it poses, is the information provided by AI safe to consume for general public? Can AI be weaponized using propaganda or disinformation? No certain answer yet.
In the picture below, you can see the ChatGPT moderation filters in action along with its political biasness.
How AI will Kill the Search Engines (Google, Bing)
AI is receiving all the media attention and rightly so. I acknowledge this technological breakthrough and its capabilities but I am afraid it is becoming another glorified bubble which temporarily has good benefits for the users in the beginning, but when the bubble pops, chaos ensues.
Majority of internet users rely on Search Engines such as Google, Bing or Baidu to find solution to their problems or read news, or learn new information. With the arrival of AI Chatbots and their increasing parameter improvements and popularity, people will eventually ditch search engines and type their query in ChatGPT or MidJourney. Googling will become thing of the past as the AI offers quicker answers even though the reliability of those answers is questionable, as fewer people do the fact checking.
When the Search Engine usage dies down and AI Chatbots usage peaks. Bad Things are going to happen.
Allow me to explain from a content creator perspective:
ChatGPT, MidJourney and similar AIs feed on publicly generated content to provide answers. There are millions of writers, artists around the world who generate their content and earn livelihood by selling their art or expertise through websites, ad revenue or freelancing. When the AI reaches its peak, millions of those writers/artists will stop producing new content because its not profitable for them to continue any further. The websites of content creators will become ghost town with little to no traffic, as casual internet users would prefer typing ChatGPT in web browser instead of Google.com.
The AI Bubble Bursts
There’s a popular saying, ‘What goes up must come down’, and this fits perfectly here. When everyone switches to the AI Chatbot and ditch search engines, the wave of freshly produced content stops, as there’s no incentive for artists to create any meaningful art. The AI’s hungry Taz Monster has no data to feed on. The ChatGPT starts losing its correctness, efficiency over time. Its output answers become outdated and irrelevant due to stalled incoming stream of data.
This AI bubble could take couple of years to fully blow and pop. During this process, browsing internet will feel repetitive, unhuman, emotionless. You would also get bombarded with disinformation and conflicting political opinions hardcoded in the AI training models. Moreover, there will be a significant industry level changes as well as career transitions by professionals/creatives. In simple terms, prepare your mind to ignore walls of meaningless Lorem Ipsum text.
Conclusion (The Good News)
We humans are social creatures by nature, to survive we need some form of social connection. Surely, AI can mimic such behavior very well but deep down its unsettling and unfulfilling for us. For example; Multiplayer games. Popular game titles such as Counter-Strike, League of Legends, Overwatch or even old-school Chess have Practice/AI mode where you can play against computer. These mods have different level of difficulties from beginner to expert, the AI player, in some cases is unbeatable too. Although online gaming can sometimes become toxic due to chat communication, we still prefer to compete against humans because it feels more lively, fun, and engaging. There’s beauty in imperfection too.
In this context, AI will reach its peak like the Dotcom bubble, then it will come crashing down and correction will occur forcing Internet giants and startups to change their priorities. State governments will create their own versions of AI and weaponize it. New/strict Copyright, Intellectual Property laws will be drafted. Search Engines will gradually become mainstream again and they will try to incentivize scholars, artists and journalists to produce content, in attempt to revive the ad revenue
… till the cycle repeats.
