AI Chatbot

By Emil Bjerg, journalist and editor of The European Business Review

“The future is already here – it’s just not evenly distributed” a quote from William Gibson goes. With ChatGPT, many users have reported a mind-blowing feeling of stepping into the future – and with free global access, for once – and for a while – the future is evenly distributed. This piece examines the impact of ChatGPT – in dialogue with ChatGPT.

ChatGPT has been the talk of the town since the launch of the publicly available beta test in early December. While it took Facebook 10 months to amass a million users and Netflix three and a half years, ChatGPT had a million users within five days. The service is so popular that you’ll more often than not see the chatbot warning about capacity issues.

ChatGPT is a generative AI, meaning that it can generate original content based on prompts given by users. But actually… what better way to introduce an AI language model than to have it present itself? According to ChatGPT, it’s “a powerful language model developed by OpenAI that can generate human-like text. It has been trained on a massive dataset of internet text, allowing it to understand and respond to a wide range of topics and questions”.

Behind ChatGPT is OpenAI, one of the most ambitious companies working with AI, that  also created the AI image generator Dall-E. Microsoft recently invested 10 billion USD in OpenAI, planning to use ChatGPT as a part of its Office Suite.

A technology full of promises and pitfalls

ChatGPT is a multi-purpose technology. You can use ChatGPT to give you personal advice, get a full-week vegan meal plan, do your homework, give you an original business idea or have it write your job application. ChatGPT even passed an MBA exam at the prestigious Wharton School of Business.

In fact, the chatbot can assist with so many errands and tasks, that we might as well define it by what it can’t do. 

ChatGPT is officially trained on data until 2021, meaning that it’s not a good sparring partner for present events. While it could have written an article similar to this – although slightly more stiff and formal – it couldn’t fact-check or evaluate the quality of it. 

Or as Forbes puts it, it lacks skills that humans excel at such as critical thinking and creativity. On the other hand, humans can’t answer complex math puzzles in milliseconds, so ChatGPT offers great opportunities for hybrid human-AI collaborations

We’ll come back to the pitfalls of ChatGPT later. First, let’s have a look at how much of an impact the new software has had already. 

ChatGPT has Google in ‘Code Red’

With ChatGPT’s growing popularity, Google’s CEO Sundar Pichai recently issued a “Code Red” in the search engine company, since ChatGPT “looked as if it could offer a new way to search for information on the internet” – and in doing so threatening Google’s business model. 

Until recently, Larry Page and Sergey Brin, the founders of Google, were no longer involved in the company’s daily operations. That changed with Pichai’s Code Red. Page and Brin have joined several meetings with executives in Google where they have reviewed Google’s artificial intelligence product strategy as well as approved plans and pitched “ideas to put more chatbot features into Google’s search engine”. According to the New York Times, “Google now intends to unveil more than 20 new products and demonstrate a version of its search engine with chatbot features this year”. 

Does ChatGPT itself think it’s a threat to Google’s business model? GPT says: “It is unlikely that ChatGPT would be considered a direct threat to Google’s business model, as the two technologies serve different purposes. While there may be some overlap between the capabilities of ChatGPT and Google’s products, it is unlikely that ChatGPT would have a significant impact on Google’s overall business model”.

ChatGPT isn’t trained on data from 2022 and 2023, mind you. Was it aware that it had acquired a million users in five days and had Google – of all companies – in a Code Red, it might have answered differently. 

So when the chatbot can cause that reaction at Google, how will it affect smaller, mortal businesses?

Generative AI and its uses cases for businesses

For most businesses ChatGPT comes with more opportunities than red flags. A recent McKinsey article foresees six departments in companies where ChatGPT and other generative AI will be impactful: marketing and sales, operations, IT/engineering, risk and legal, HR and R&D. 

Some of the more advanced use cases for AI McKinsey foresees are answering questions based on large amounts of legal documents and writing marketing copy. The general pattern of the use cases McKinsey suggests has to do with summarization, task automation, and creation of data, text, and code. 

When given the prompt “How can ChatGPT and other generative AI impact businesses?”, ChatGPT mentions content creation (“Generative AI can be used to generate text, images, and videos”) and customer service (“AI-powered chatbots can provide 24/7 customer service”). 

It also mentions personalization (“Generative AI can be used to personalize product recommendations”), product design (“Generative AI can be used to design new products, optimize manufacturing processes, and predict product failures”), and risk management: (“Generative AI can be used to identify and mitigate risks in areas such as finance, healthcare, and cybersecurity”).

When prompted to provide more recommendations, ChatGPT provides six other use cases – among them translation, research and development, and fraud detection – each one valuable in the right context. 

To most businesses, the question is not if ChatGPT can assist them. It’s whether executives and employees can spot how it can help and how to give it good prompts. 

However, there are also issues to be wary of when it comes to generative AI, both from a business- and a societal perspective. 

With ChatGPT, we’re all ‘prosumers’

“If it’s free, you’re the product” a saying attributed to Andrew Lewis goes. With ChatGPT, we’re all ‘prosumers’, meaning we simultaneously produce and consume the product. 

In other words, the reason ChatGPT – an overall valuable tool that OpenAI could charge for – is free is because it’s in the testing phase. When students use the chatbot to do their homework or when it passes that MBA exam at Wharton School of Business, it’s not just flexing skills. It’s training. The team behind the chatbot uses the prompts and replies to test the wide range of reasonable and unreasonable tasks the chatbot receives and uses those to regulate and optimize functionalities. 

After the testing phase, we can expect to pay for it, as we do with most other things that make life easier. Users already report having been offered ‘pro’ access for 42 USD a month

Awareness of bias and misinformation

One issue to be wary of when using ChatGPT is the risk of bias and misinformation. The GPT part of ChatGPT stands for ‘Generative Pre-trained Transformer’. That means that the answers it gives are no better than the statistical average of the data created by humans that the chatbot is trained on. Or as ChatGPT puts it, it may “have learned biases present in the data. This can lead to the model generating biased or offensive responses”. 

Just like the chatbot can produce biased text, it can also produce misinformation. Currently, businesses need to fact-check everything – even calculations of math puzzles. ChatGPT feels convincing – even when it’s wrong.

ChatGPT and the fear of automation

ChatGPT also revokes old but ever-relevant discussions around the risk of automation taking jobs. The industrial revolution replaced physical workers in factories and farms, and a couple of centuries later digitalization has automated countless bureaucratic tasks. 

This time it’s the creative knowledge workers – coders, designers, writers, and more – that might wonder if the latest technological leap will take their jobs. 

So far, the pattern has been that when jobs disappear to automation, new jobs emerge. That might be the case with this breakthrough as well. As Forbes puts it, ChatGPT lacks abilities like strategic decision-making and the aforementioned critical thinking and creativity – central traits for knowledge workers. Instead, they see that chatbots can supplement customer service staff “allowing people to focus on more complex and more ‘human’ tasks”. The Atlantic is less optimistic on behalf of the white-collar workforce, stating that it is likely that AI in the coming five years “will begin to reduce employment for college-educated workers”. 

What does ChatGPT itself say about the matter? Asked “Can ChatGPT and other generative AI result in a loss of jobs?” it replies: “Yes, it is possible that the use of generative AI such as ChatGPT could result in a loss of jobs, particularly in fields where the technology can be used to automate tasks that were previously done by humans. However, it is also possible that the use of such technology could lead to the creation of new jobs and industries. It is important to consider the potential impact of these technologies on the workforce and take steps to mitigate any negative effects”.

Whether chatbots will steal white-collar jobs on a massive scale is still to be settled. The aforementioned article from The Atlantic gives valuable advice that’s worth passing on to our readers: “It will be important for individuals to stay up to date on the latest developments in AI and to consider how their skills and expertise can be leveraged in a world where machines are increasingly able to perform many tasks”.

LEAVE A REPLY

Please enter your comment!
Please enter your name here