Who Asked for Another Chatbot?

Authored by Christopher Barnes, Vice President of Engineering, Data & Analytics, insightsoftware

This may be an unpopular opinion but I, for one, was not looking for another chatbot or automated assistant; I’m all set with the number of robotic helpers and can frankly live without more.

They have been around for years. You find them on every e-commerce website, every customer support site, every answering service at a doctor’s surgery or hotel booking line… not to mention Siri or Alexa listening in the background in every kitchen, living room, and bedroom around the world. Being a professional nerd, I have adopted and used what has passed for AI assistants up to this point – the interactions have been unremarkable, narrow in their intent, and hit-and-miss in terms of the results. I have lost count of the number of times I have shouted down the phone, “just put me through to a human!”

The generative AI revolution is in full swing, and the delivery method of choice is your free-form AI-powered chat assistant. Whatever you may be doing, whatever you want to ask, whatever you want to accomplish, the AI can help you, right?

Not always the case!

Undoubtedly, there is a step change in the usefulness of the interactions, but the user is typically left alone to figure things out for themselves. This is because natural language isn’t all that natural to everyone, and context isn’t the same for everyone. Plus, Gen AI bots are neither deterministic nor always telling the truth. And that’s before you consider all the spurious or opinionated data that they’ve been trained on!

Chatbots tend to be easily manipulated and wide open to misuse, giving the technology a mixed reception to say the least. I’m sure we’ve all seen this notorious example:

But in the grand generative AI experiment, we are all just guinea pigs, and this is just the beginning. The early adopters and the beta programs are being run to test what works and what doesn’t and to test the tolerances of society. We get the extraordinary successes and the spectacular failures in real-time – the failures seem to get more airtime than the successes, but there are plenty of both surfaced in the media.

 ChatGPT: Why we’re still smarter than machines – BBC

Why Google’s ‘woke’ AI problem won’t be an easy fix

AI: Google restricts Gemini chatbot election answers

Nope. It is definitely not just hype. The good here is VERY good and likely represents a revolution in the way that we live and work – perhaps to the same degree as when Tim Berners-Lee brought us the World Wide Web. We are at that point on the hype curve where every software company in the world is scrambling to incorporate AI wherever they can even though no one is quite sure which direction it will take us. But while there is an upside to getting it right and a definite risk to getting it wrong, there is also a cost to not even being on the playing field.

However, the current state of AI assistants and chatbots is like using a rotary phone to access the internet. They’re clunky interfaces for powerful technology. That’s where purpose-built, context-aware AI comes in. Imagine an AI that understands your needs, can access and process information from various sources, and guides you intelligently. This is the true potential of AI; embedded into users’ daily routines. A network where the AI can constantly learn, be aware of its surroundings (system aware), and be driven by a specific purpose (like improving your workflow). This is how we unlock the trusted and powerful AI of the future, not through clunky chatbots but through intelligent assistants built on open data.

It is always dangerous to try to predict what is to come, particularly with the advances of technology:

• “Before man reaches the moon, your mail will be delivered in hours from New York to Australia by guided missiles. We stand right at the brink of an era of rocket-powered mail.” Arthur Summerfield, U.S. Postmaster General; 1959.
• “Mobile phones will absolutely never replace the wired telephone”. Marty Cooper, inventor of the mobile phone, 1981.
• “The internet will fade away because most people have nothing to say to each other. By 2005 it will be clear that the internet’s impact on the global economy has been no greater than the fax machine.” Paul Krugman, renowned Economist, 1998.

But, if not chatbots, then what does the future of generative AI look like? The new wave of AI technology is not going to replace people, and there are not going to be teams of AI bots doing the work currently done by teams of people. It’s unlikely that the AI of the near future will be able to replace the ingenuity, creativity, and imagination of the current workforce. However, if we are honest with ourselves, there are parts of almost anyone’s job in almost any industry that don’t require imagination, creativity, panache, or ingenuity. In fact, in the corporate world of emails, documents, PowerPoints and virtual communications, you could argue that quite a lot of the grind can be picked up by AI.

AI has the potential to supercharge modern-day corporate employees and maximize what we do best and where we differentiate ourselves from the machines. This will happen as organizations start to embrace large language models and generative AI with more precision and less breadth. As the industry zooms in on the use cases that really make a difference, they will start to specialize in the areas that yield the biggest productivity gains. Just as no middle schooler is sifting through the Encyclopedia Britannica for their geography homework when they can just “Google It,” it won’t be long before the vernacular changes once more. Perhaps we will revert back to “asking Jeeves” or some other evolution of the same!

This focus on specialization through AI integration shows us a potential future for customer benefits in software development. While you might currently find chatbots frustrating, advancements in generative AI could lead to a more nuanced approach. Imagine context-aware AI that acts as an intelligent bridge between your development team and your customers. By understanding customer needs and feeding them directly into the software creation process, this AI could foster a more symbiotic relationship. However, it’s important to acknowledge that, similar to the tentative shift from physical encyclopedias to search engines, this future of generative AI in software development is still under development. The potential for a more collaborative customer experience hinges on the industry’s ability to refine and specialize these AI tools.

As software developers, software product managers, and designers, what should we be striving for after the initial chatbot scramble starts to lose its shine? Perhaps we should be thinking about AI advancement from the point of view of the consumer:

• What do the users do every day?
• What do they do that takes time?
• What do they do that drains productivity?
• Which of those activities fall into the natural proclivities of generative AI?

The software industry’s test is to embed generative AI so that it sits in the background, offers recommendations, provides timely analysis, and prompts users to make better decisions when they need an injection of creativity or inspiration. AI needs to gather research, summarize communications, and surface anomalies and trends to drive faster and better-informed decision-making.

So, who needs another chatbot? I guess we do.