A.I. is ALREADY Replacing Jobs in the US

Author(s): Aditya Anil

Originally published on Towards AI.

A new survey shows that AI is ALREADY being employed by many business leaders based in the US.

I’m not sure if you’ve heard, but AI is taking away jobs in the US right now. Yes, you read that right.

The technology that was supposed to make our lives easier and better is now making a ‘life’ for itself.

Well, it turns out that it’s also making some of us redundant and obsolete. And this was clear from a survey by Resumebuilder[dot]com

Business Leaders Favor AI: Stats shows

According to this survey, almost half of US companies have already implemented ChatGPT. And as a result, it has replaced workers at their companies. And many of these works, as you may have guessed, are related to creative fields or (mostly) non-technical fields.

Seems you no longer need ‘actual’ intelligence if artificial intelligence has outworked you, right? Talk about the quality and quantity comparison, AI offers both. Billions of calculations and many fine tunings seem to be enough to work for any MNC.

Mind you, ChatGPT is not your average chatbot. It’s built on top of OpenAI’s GPT-3 family of large language models, which means it can generate coherent and fluent text on any topic. Although many may not be perfect, it’s not bad after all. How do I know? Because we saw in front of our eyes that this seemingly small chatbot can write code, create poems and content, handle customer service, and even summarize meetings and papers.

“There is a lot of excitement regarding the use of ChatGPT,” Resumebuilder.com’s Chief Career Advisor, Stacie Haller, said in a statement. The survey that we are discussing here was undertaken by them. “Workers need to surely be thinking of how it may affect the responsibilities of their current job. The results of this survey show that employers are looking to streamline some job responsibilities using ChatGPT”.

It won’t take another survey for you to know why employers favor these not-fully-reliable robots to take up a job. What is the obvious answer? Scale and Economical advantages.

Scale since AI chatbots like such can be optimized to do any task at any level. And with some fine-tuning, it can generate thousands of worthy content that can look ‘appealing’ to the consumer. And as for the economic reason, you don’t have to pay for it. Apart from initial setup cost (which is not too much if you outsource it) and maintenance, these chatbots don’t take much “salary” to do their work. When was the last time that you paid your virtual assistants like Siri or Google or Alexa?

What do the Numbers say?

Most of the jobs that these robots take up themselves have very high competition. The survey showed that 66 percent of US companies use ChatGPT for writing code, 58 percent for creating content, 57 percent for customer service, and 52 percent for meeting summaries and other papers.

Oh, and what do the business leaders think of ChatGPT’s work? Well, ironically, they love it. Fifty-five percent say the quality of work produced by ChatGPT is ‘excellent’, while 34 percent say it’s ‘very good’. Only 11 percent say it’s ‘fair’ or ‘poor’. That’s insane.

Numbers don’t lie unless it’s in statistics. And since this was a survey done on 1000 business leaders, I am a bit skeptical about the result. Sure, ChatGPT could do ‘excellent’ in paperwork tasks and writing content, but the truth is that these robots are ‘poor’ when it comes to research-based studies or calculation-based work.

Ray of Hope

Are you scared yet? No, you shouldn’t. And here is a solid reason why.

‘These robots are good at math,’ you may say. Well, yes, for solving elementary equations. That's a bold statement to say. It's simple, gives a word problem to it, and it won’t take you more than 10 or 15 tries to see that this machine fails. Arizona State University found ChatGPT was terrible at math. It scored with below 60% accuracy in the 1,000 mathematical word problems given to it by the professor. And we are talking about accuracy, not speed. It doesn’t matter if it can spit out the answers in a few seconds if, ultimately, it gives the wrong result.

So now I am asking the same question that both you and I once asked when we got to know about the movies regarding AI. Is this such a good idea that our jobs are being taken over by ‘unpredictable’ robots? Sure, humans are unpredictable, too, I get it. But they can be tamed, punished, and corrected. And we know how to go about it in the case of humans. But in the case of robots, we are blinded by the algorithms and invisible technical bugs.

Is ChatGPT trustworthy and reliable? No, at least not fully. What if it makes mistakes or gives wrong answers? Then we are doomed. What if it has biases or hidden agendas? It already does. What if it goes rogue or gets hacked? Welcome to the terminator.

While we are debating on the above questions (not just hypothetical questions, but serious ones) Sam Altman, the founder of ChatGPT and CEO of OpenAI, has already warned users against relying on the AI chatbot for “anything important”. He also expressed worry about the risks posed by artificial intelligence in general.

Unreliability of present unreliable chatbots

But ok. The answers above were filled with ‘hope’ and ‘scepticism’ (and ego, maybe). But are there any quality factors that will ensure that these robots won’t take up ‘many’ of our jobs?

Well, my friend, yes.

What are the two biggest yet most common money-making industries, according to you? Probably Oil and banks. And AI failed miserably in these fields. Big bank employees say they’re running into issues using ChatGPT to make their jobs ‘easier’. (but first of all, is my money being looked over by a robot ?!)

Many traders told Bloomberg they had used the popular AI chatbot to aid in their jobs. But to their dismay, it had several limitations, such as having to spend a lot of time fact-checking and removing out-of-date information. If you don’t know, ChatGPT doesn’t have all the latest information like our traditional search.

Therefore, when one oil trader found there was out-of-date information in the crude market outlook (written by ChatGPT), they that had change it. And since these AI machines generate A LOT of data, it takes A LOT of time to fact-check it.

A bank found that ChatGPT compiled an overview of a client in less time than an internet search would take. Cool, but they ran into another problem. The info they gathered than had to be cross-checked. This cycle of cross-checking paradoxically increased our time.

And from a logical standpoint, if you need to crosscheck to be done, you need ‘humans’ to be present there. And why not, since all that happens in business is that one human works to provide for another fellow human and then later gains from it? So while it may seem these bots are highly capable of taking our jobs, they cannot fully rely.

“When we talk of high-accuracy tasks, it is worth mentioning that ChatGPT sometimes hallucinates and can generate answers that are seemingly convincing, but are actually wrong,” Morgan Stanley analysts said last week. And big corporations do take these false positives seriously.

JP Morgan has already put restrictions against the use of ChatGPT among traders because it’s worried that sharing sensitive financial information with the tool will ring regulators’ alarm bells. Wall Street banks Citigroup and Goldman Sachs have also blocked the chatbot on third-party software.

TCS in India has understood the AI takeover better. TCS has already said that generative artificial intelligence platforms like ChatGPT will make an “AI co-worker” and not lead to job losses. This is a positive and reassuring way to see the whole situation. AI being your assistant and doing the tough job is much better, instead of AI taking the tough job and making our lives even more tougher.

“It will be a co-worker. It will be a co-worker, and that co-worker will take time for them to understand the context of the customer,” Milind Lakkad, the chief human resources officer (CHRO) of TCS, said in an interview with news agency PTI recently.

Conclusion

So hopefully, I am pretty sure that there won’t be a complete hostile takeover of our jobs by AI. Why? AI is created by humans after all, and while it can learn on its own, the algorithm is also made by humans. And human makes mistakes, which we have to be cautious about. There will always be some downsides to any new IT revolution. In this case, AI can’t be fully relied on. And while we can have a huge ‘quantity’ of posts with us using AI, we will always favor ‘quality’ in the long run.

Otherwise, we would have to spend twice the amount of time to ‘fact check’ our own choices.

And that is it for today. And as long as AI doesn’t take over my work, see you in the next.

If you liked the article above, do check out my weekly newsletter — https://creativeblock.substack.com where I share AI and Tech related articles

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI