We are in an AI hype right now, but the AI winter will come again….
Everyone is talking about AI these days — especially with ChatGPT and Bard from Microsoft and Google dominating the playfield. Even people who don’t conventionally like change have started to understand the value which AI can bring to their organisations. And of course, Andrew Ng rightly said many years ago that “AI is the new electricity”.
With that, I am hearing about people fearing about AI — will AI take our jobs, will it replace us in every walk of life, will humans not be needed and lots of similar questions. And these questioners are not just humans who are currently in jobs that ChatGPT or Bard have shown promise to automate (e.g. marketing and social media)— these include humans who are in the domain of IT itself — ranging from data scientists, machine learning engineers, website developers to software developers and even full-stack developers. Everyone has this fear. Fear of missing out (FOMO) if you are not into AI right now, or fear of your work being taken over by AI if it continues to grow at an exponential rate as of today.
Not many people are aware of the AI Winter that prevailed from 1974–1980. Wikipedia defines AI Winter as “a period of reduced funding and interest in artificial intelligence research”. When AI was in itself coined as a term in the 1950s, there was a period of increased interest and excitement in its potential to disrupt organisations and change lives. However, this was not permanent. This excitement dampened down and nearly died during the period of AI Winter. What would happen if you are using a technology for years to do similar types of actions — it would become monotonous and you would stop to value the power of the technology. Same is true for AI — while right now in 2023, because we are actually seeing such large language models like ChatGPT for the first time in our life, they seem amazing to us. It feels like they can do anything and everything. And it is rightly so, they are indeed amazing, powerful and have huge potential. But let us not forget that as we continue to use such models for hours every day for 365 days a year, we will start to understand the monotonicity of the models; their limited capabilities; their usual ways of responding to questions and so on. These models would become so usual for us soon that we would not even think of their power and potential in the next few years but consider them as obvious and foundational.
That’s where the direct response to the fear comes out — AI will not replace us, but it will replace our ways of working. Yann LeCunn has rightly said:
“Calm down. Human-level AI isn’t here yet, And when it comes, it will not want to dominate humanity. Even among humans, it is not the smartest who want to dominate others and be the chief. We have countless examples on the international political scene.”
Let us not get too excited about this new technology — it is great but not even close to singularity or passing the Turing test. It is far, far away from there.
AI is only going to replace us if we don’t know how to use it. If you are a data scientist, you need to give the right prompts to the AI to make it generate the right code to develop your AI models or visualise your data. If you are a website developer, you need to give the right prompts to the AI to make it generate HTML and CSS code for the website to look exactly like what you desire. If you are a marketer, you need to ask the AI to generate the right content that has a human touch to it. In summary, you need to question the AI — not just to receive answers, but question the answers in themselves. Who checks that the code generated is correct, the marketing content is eye-catching and the data you have is trustworthy? None other than us humans. As Ankur Warikoo rightly puts:
“If you know how to ask questions, the machine will not replace you. It will exist because of you”.
This is the right time to get into AI, play with AI tools like ChatGPT and Bard, get their help and support in making our everyday lives at home and work easier. This specially is integral if you don’t want to resonate around the FOMO. However, we should be wary of an AI winter that will inadvertently happen in the next few years. When will it happen? No one knows or can predict — but it will surely happen. The excitement around AI and its potential will seem to have died, and there will only be few people in the room of an organisation talking about AI at that point. There will be little money flowing into AI startups. Companies will be pursuing limited investment in AI at that point. Maybe academic institutions will also see less students interested in studying AI at that point!
As we continue to use these tools more often, we will see the false information it can utter from time to time, their lack of ethics and moral values, inability to have a human touch in activities that so conventionally are built on human touch (e.g. marketing, HR etc.) and so on. As a notable point for individuals already in IT — you will still need to be around at that time as everyone would realise that it is impossible to depend on an AI to build softwares and applications (or even AI to write AI!) in safety-critical applications, specially in industries where even the slightest mistake by our good old AI assistant can potentially lead to a complete foiling of a business. For every human out there — your roles will continue to become even more important than before. AI will not take over your jobs, but you will take over the AI. Just remember to ask the right questions and use the tools for supporting you, instead of using them in every walk of life.
Our human brains are amazing — and the AI models will most-likely never reach the ability to think and understand things as we can. Yes, we do see news and hype circulating around the internet on how these AI tools can easily solve complex business problems spanning diverse areas — but how much have we even used these models yet in real-world tasks? We are just at the very beginning. And as we actually try and adapt these tools to real-world tasks, we will soon realise that we are more intelligent as humans than the AI for a “lot” of things.
Of course, as it happened in the previous AI winter during 1970–80s, there will be a bounce-back and a new era of AI will usher again — by tackling/mitigating the pitfalls that the current AI tools bring with them and going above and beyond their current capabilities. This is a never-ending cycle of growth, downfall and again back to growth.
Without being too negative about the future of AI, this is the right time to invest in AI — invest in learning more about AI, invest in using AI more often in your businesses and understanding what it can and cannot do, investing in AI startups/companies (if you are a financial investor), invest in upskilling your future generations on AI and so on.
This investment will be a boon when the next AI winter comes in the next few years — as having already learnt so much about the strengths and weaknesses of current AI models, researchers and academics, businesspersons and managers, data scientists and analysts, HRs and marketers, sales and finance professionals, legal professionals etc. and many more would be able to help support a new beginning for AI. My grandfather used to say:
“The beginning of an end is the end of a beginning”.
We should never think that the next AI winter would mean the death of AI, but it would be an opportunity to bounce-back, being more resilient and make a bigger jump on what AI can do there! Let us embrace AI as a technology, but not fear it. If we invest in a future that brings together humans and AI, it would be a bright future — bright for governments, academia and industry as well as for humans in themselves.
I asked Bard to summarise this article (written by me). Here is its response and a very nice summary of Bard’s main takeaways from this article, hope you will also have had some takeaway from this article:
We are currently in an AI hype cycle, but an AI winter will come again.
- The AI winter of 1974–1980 was a period of reduced funding and interest in artificial intelligence research. This was due to a number of factors, including the failure of early AI systems to live up to expectations, the lack of progress in areas such as machine learning, and the ethical concerns that were raised about AI.
- The next AI winter could be triggered by a number of factors, such as the failure of current AI systems to live up to expectations, the lack of progress in areas such as natural language processing, or the ethical concerns that are raised about AI.
- The impact of the next AI winter could be significant, as it could lead to layoffs, a decline in innovation, and a loss of interest in the field. However, it is important to remember that the AI winter is a cyclical phenomenon, and that interest in AI will eventually return.
