2023: The Year Of Artificial Intelligence
How long does it take businesses to acquire one million users? Twitter took 24 months to hit that number, Facebook took ten, Spotify five, Instagram two and a half and Path a mere two weeks. In December 2022, an artificial intelligence (AI) based chatbot capable of understanding– and responding in – natural language, called ChatGPT, broke all records.
It took mere five days to acquire one million users. Here is why: it is smart and intelligent and anyone could use it without a doctorate in AI. Users have described it as “frighteningly good.” It holds the promise of making 2023 ‘The Year of AI.’
You ask the Chat Generative Pretrained Transformer (ChatGPT) a question, like “why is Newton's first law incomplete?” and it provides examples where Newton’s first law of motion fails. The response is lucid and concise. It is better than what most humans, even high school physics teachers, would be hard-pressed to come up with.
ChatGPT is doing amazing things: helping write the music and lyrics for songs, assisting with holiday shopping, drafting college applications, making up stories, playing games, writing code, and even doing some distasteful racial profiling.
What AI holds for 2023 is an interesting question to explore— and a necessary one for top executives across industries.
AI, without any fanfare, has become the go-to technology for every business. Want to fight financial crime? Thinking of planning better, smarter, more sustainable cities? Wish to create perfumes without touching a single ingredient and personalise them for a market of one? Determined to replace thousands of years of traditional craftsmanship that go into whisky making? Predict injuries in athletes and sportspersons? Provide speech-to-speech translation? Want to pollinate crops but don’t have bees to do the job? Want to improve waste recycling? Just use AI. Greg Brockman, President and Co-founder of OpenAI, which developed ChatGPT, says, “2023 will make 2022 look like a sleepy year for AI advancement and adoption”. It is difficult not to agree with Brockman.
Most new technologies capture our imagination with a bang. Mobile, Cloud, Big Data, and APIs are just some from the past that stayed in the headlines for years. However, there was nothing to indicate AI’s sudden and swift rise.
Yes, there has been much talk about AI, but no serious money behind it. Businesses have been exploring AI but have yet to be able to scale their projects or extract the promised ROI. The reasons are varied—some lack the right tools, others don’t have the right talent, some have experienced unpleasant project cost overruns, and others suffer from poor data quality or have security apprehensions.
An examination of data shows that AI, as a technology, has been flying under the radar. Funding for AI start-ups has been painfully low and slow. Worldwide, investments in AI startups dropped from USD 72 billion in 2018 to USD 65 billion in 2019, hovering around the same mark at USD 69.5 billion in 2020 before rising to USD 111.4 billion in 2021 and then plunging to USD 47.5 billion in 2022 (projected).
In fact, of the 1,211 unicorns worldwide in 2022, just 90 were from the AI space. Compare this with 258 in fintech, 108 in e-commerce and 129 in the healthcare industry and it becomes clear that AI still needs to create a significant blip on VC radars. So, what makes AI such a hot technology for 2023? There are two reasons why AI is going to take 2023 by storm. Let’s explore those two reasons.
First, large data sets and computational capabilities to train AI were not readily available. Training AI was an expensive job. Not any longer. The cost of training image classification systems has dropped by 63.6 per cent, while training time has improved by 94.4 per cent since 2018.
Computational power, too, has grown at a ferocious pace. In 2016, AlphaGo, the first computer program to defeat a professional human Go player, was trained on 1.9 million peta floating-point operations per second (PetaFLOPS— a measure of training computation, one FLOP being one addition, subtraction, multiplication, or division of two decimal numbers). Minerva, in 2022, was trained on 2.7 billion PetaFLOPS to solve mathematical problems at a college level.
These capabilities have made it possible for AI to exceed human benchmarks. Between 2017 and 2020, AI exceeded human capabilities for handwriting recognition, speech recognition, image recognition, reading comprehension, and language understanding.
Falling costs and rising capabilities will force industry leaders to turn their attention to AI. One of the best ways to gauge the tremendous strides AI is making is to understand that the technology has gone from image recognition to image creation, from transcribing and mining data to creating new artefacts like code for applications and SQL queries.
The second reason AI will take 2023 by storm is that NLP interfaces give everyone a chance to use AI without the need for training. NLP is a breakthrough. No matter what language you speak, you can fill the gap between communication and computer understanding. Siri and Alexa are examples of NLP in everyday use.
This brings us to the critical changes we must prepare for in 2023. The former director of AI at Tesla, Andrej Karpathy, says that AI is already writing ~80 per cent of his code with ~80 per cent accuracy. He says, “I don't even really code, I prompt & edit.” This provides us with an important insight into the future of programmers.
Now, instead of spending time writing a function or researching how to do it, programmers will need to know how to write the correct prompts. AI will spew the code, and programmers will review it. We can expect this to happen across the board.
For example, information workers of all shades will not need to create content; they will need to review what their AI app produces and improve it. AI will be used to transcribe and summarize meetings, create the initial set of ideas for projects, draft social media posts and emails, etc.
Then humans will review the work AI puts in, fine-tune it and use it. In effect, the concept of the intern is about to end. AI will be the new intern.
You can already see where this is headed. Professionals from all industries will hasten to power their careers by acquiring No-Code MBAs that allow them to use AI with greater felicity. How to use AI will be the hot new skill to acquire in 2023.
