Episode 6: The Chatbot Epidemic - The Modern IT Plague

Transcription:

Hello, and welcome to Speedtalks, the podcast that drives deep into the latest trends in technology, especially how it’s reshaping the financial industry. I’m Tim Olszewski, and today we are tackling a topic that’s both groundbreaking and unselling – the chatbot pandemic, the modern AI plague.

Introduction

In today’s episode, we are going to explore the rays of chatbots. These ones niche tools have now become ubiquitous in our daily lives and the business landscape. From customer service to marketing, chatbots are everywhere, promising efficiency and cost savings. But, as with any rapid technological advancement, there are significant risks and challenges that can’t be ignored. We’ll start by outlining the explosive growth of AI projects in recent years, then delve into why chatbots have become one of the most popular use cases.

But it’s not all good news. We’ll also examine the darker side of this technology, the security, vulnerabilities, privacy concerns and real-world consequences of relying too heavily on AI-driven chatbots. So stick around as we dive into the good, the bad and the downright dangerous aspects of the chatbot revolution.

AI Chatbots Market Overview

Now, let’s begin with the bigger picture. In 2023 alone, the industry produced 51 notable machine-learning models, more than double that of the previous year. This surge in AI development isn’t just a trend – it’s a transformation, with the number of AI-related projects on GitHub skyrocketing from just 845 in 2011 to approximately 1.8 million in 2023. This growth is reflected in the chatbot market, too. By 2023, the global AI chatbot market was valued at USD 6.4 billion and is expected to soar to USD 66.6 billion by 2033, growing at a staggering CAGR of 26.4%. With 47% of organisations planning to use chatbots for customer care and over 50% of enterprises expected to invest more in chatbot development than traditional mobile apps, it’s clear that chatbots are becoming an integral part of business strategies worldwide.

Real-world Examples

But what does it look like in practice? Let’s explore some of the real-world examples of companies that have successfully implemented chatbots.

Bank of America – Erica

One of the most recognised chatbot implementations in the financial sector is Erica, Bank of America’s virtual financial assistant. Launched in 2018, Erica helps customers with everything from checking balances to making payments and providing personalised financial advice. By 2022, Erica has already surpassed 1 billion interactions with customers, demonstrating how significant the impact chatbots can have on customer engagement and service efficiency.

Domino’s Pizza – Dom

Another example is Domino’s. Domino’s introduced its chatbot Dom to, so to speak, revolutionise the ordering process. Customers can place orders via voice or text, with Dom making it easier and faster to get their favourite pizza delivered. And the chatbot is integrated across multiple platforms, including Facebook Messenger, Google Home, and Amazon Echo, showing the versatility of chatbots in enhancing customer convenience.

Lyft – Lyft Chatbot

Another example is Lyft, the ride-sharing company, which has integrated a chatbot within its app to allow users to book rides without needing to open the full application. The chatbot provides real-time updates on drivers’ locations, estimated time of arrival and more. This implementation was expected to make the process of booking and managing rides more seamless for users. But the reality is very often a little bit different from expectations.

AI Chatbots in Business

Anyway, these examples highlight how chatbots are not just, you know, a novelty, but a practical tool for enhancing customer service, driving sales, and improving overall business efficiency. The application of chatbots extended across industries, from retail and banking to food services and transportation. It’s everywhere. They have become so prevalent that nearly every industry is finding ways to leverage chatbots. The question is, is this something that those companies have integrated within their sales and business strategy? Is it always the case that those chatbots are in the right place in the company’s business model? Well, that’s a very good question.

The widespread adoption is largely driven by the tangible benefits chatbots offer. For instance, they are expected to handle between 75-90% of customer queries by 2024, potentially reducing customer support costs by up to 50%. Moreover, businesses report significant improvements in efficiency and customer satisfaction, as chatbots can operate 24/7, providing immediate responses without the need for human intervention.

But another question is, do we have a problem with cherrypicking? Very often, companies provide reports showcasing tangible benefits that their chatbots offer, but those benefits very often are also strictly related to marketing policies or marketing strategies because chatbots are not only a tool for reducing costs but, very often, a tool to attract people, to attract more leads, attract people to use certain functionalities in our mobile or web apps just because there is an AI agent behind the screen looking forward to service anything that we want.

The Popularity of Chatbots

As we continue to see these developments, it’s clear that chatbots are not just a passing trend but a cornerstone of modern customer service and business operations. Their ability to improve efficiency, reduce costs and enhance customer engagement makes them a vital tool in today’s tech-driven world.
So why are chatbots so popular? For one, they offer a significant return on investment. Or actually, they’re expected to. AI has been shown to decrease costs and increase revenues, with 42% of organisations reporting cost reductions and 59% seeing revenue increases after implementing AI, including chatbots.

But if we dive a little bit deeper into those numbers, we will unveil the sad truth about those high, skyrocketing numbers of cost reductions. Because while there’s 59% of companies that see revenue increase after implementing AI or 42% of organisations report cost reductions, we still have numbers saying that over 80% of all AI projects implemented by 2023 in large enterprise organisations fail after the first year after the implementation. So, if we take into account that number and we multiply it by 42% of organisations reporting cost reductions, we will be able to see how those cost reductions can extrapolate into the further horizon.

Chatbots are expected to handle between 75% up to 90% of customer inquiries by 2024, and they say the potential to reduce customer support costs is up to 50%. Well, if we would like to base our assumptions and take a look into the future based on those numbers, it will be easy to see why businesses are eager to adopt this technology. But there is a catch. So, let’s dive into some of the major challenges that come with the rise of chatbots.

The Risks and Challenges

Now, as exciting as this technology is, because I’m absolutely not trying to say that chatbots are not worth their time, this technology is, in fact, very, very exciting. It’s important to understand that it’s not all smooth sailing. In fact, there are quite a few significant hurdles that businesses need to be aware of.

High Costs with Uncertain Returns

First of all, let’s talk about costs. Companies are pouring massive amounts of money into AI. $1 trillion is expected to be spent on AI-related capital expenditures over the next few years. But here’s a kicker: we’re not seeing the returns we would expect. One of the prominent examples is IBM’s Watson. Watson was heavily marked as a game changer for healthcare, promising to revolutionise medical diagnosis and treatment recommendations. IBM has invested billions of dollars into this project, but the reality didn’t match the hype.

Watson struggled to integrate with the existing medical systems and often provided unreliable treatment suggestions. In fact, after several high-profile failures, including a partnership with MD Anderson Cancer Centre that was ultimately scrapped, IBM had to scale back its ambitions. This is a classic example of how high costs can be for an AI implementation and how difficult it is to get a return on those investments. So it’s not that simple as it is, right? It’s even though that 40-42 % of companies report cost reductions. Well, they were able to achieve cost reductions. The real question is, for how long?

Security Vulnerabilities and Privacy Concerns

Another example is obviously security matters. So security and privacy are large, huge concerns, and we’ve got some pretty stark examples to illustrate this. A major European bank, for example, implemented a chatbot to handle customers’ inquiries. They hoped to streamline their customer service operations. But shortly after launch, they discovered that the chatbot could be manipulated through specific query phrasing to reveal confidential customer information. This wasn’t just a minor glitch—this was a major breach of customer trust and data security.

And then there’s the privacy issue. According to a recent study by the Mozilla Foundation, some AI chatbots, especially those in the romantic AI space, collect extensive personal data—far more than necessary. For example, the chatbot Crushon.AI was found to be collecting sensitive personal and health-related information, such as sexual health data and medication usage, without adequate safeguards. This kind of overreach puts users at serious risk and highlights the privacy challenges that come with these technologies.

Power and Infrastructure Constraints

Another issue is infrastructure. There’s a huge strain on power and infrastructure, and it’s not a secret that AI requires a ton of energy to run, particularly when it comes to the data centres that support these technologies. Especially when we use API to communicate with LLMS that operate within our chatbots. In the US, data centres are already consuming about 2% of the country’s total electricity, and this number is expected to rise sharply as AI becomes more widespread. The problem is that the US power grid hasn’t seen substantial upgrades in nearly two decades, and it’s already struggling to keep up with current demand.

The same situation is true in most of the countries in Europe. If we don’t invest in upgrading our infrastructure, we could face a very significant power crunch that would slow down air growth across the board. For example, a state of Virginia, a high data centre hub, is already expecting stress on its power grid with the warnings that current infrastructure may not support future AI-driven growth. This isn’t just a theoretical concern; it’s a real, pressing issue that needs to be addressed.

Regulatory and Ethical Challenges

Finally, we can’t forget about the regulatory and ethical challenges, especially in the face of the AI Act. The European AI Act is a step in the right direction and, at least in most public opinion, aims to set clear rules for AI use. But regulations often lag behind the pace of technological advancement. And there is a real, you know, problem with the ethical side. Microsoft’s chatbot, Tay, is a perfect example here.

Tay was designed to interact with people on Twitter and learn from those interactions. But within 24 hours of its launch, Tay was manipulated by users into tweeting offensive and inappropriate content. Microsoft had to shut Tay down immediately. This incident highlights just how tricky it is to control and regulate AI in real-world, uncontrolled environments. It also raises questions about how we manage the ethical implications of AI, especially when it comes to preventing misuse and protecting users.

Conclusions

Finally, the question is how we create new projects and implement new AI technologies and functionalities in the face of regulatory concerns and AI governance needs. All these challenges really highlight the need for a cautious approach when it comes to adopting AI and chatbots. The potential benefits are huge, but so are the risks. Companies need to weigh those carefully and not just jump on the AI bandwagon without fully understanding the long-term implications.

To wrap up, the chatbot epidemic is both a marvel and a cautionary tale of modern IT. While the potential for innovation and efficiency is enormous, the risks are equally significant. And on that note, I’m excited to announce that in an upcoming episode, we’ll be diving deep into the AI act and its implications for the tech industry, so make sure to subscribe and stay tuned for more insights. Thanks for listening to Speedtalks. If you found this episode insightful, don’t forget to share it with your network. Until the next time, stay curious and stay informed.

This has been another episode of Speedtalks where we keep you up to date on everything tech in the financial world. I’m Tim Olszewski and remember – the future of technology is as as exciting as it is unpredictable. See you in the next episode.