There’s been a lot of talk about artificial intelligence for a while now, and until recently, for most of us it meant images of robotic machines operating equipment in factories or milking cows on farms. Sure, that’s how it started, but it’s come a long way from there. We now have robots managing robots and this is only likely to increase, and increase rapidly, given the leaps and bounds we’re experiencing in technology.

Chatbots

For some time now we’ve had chat bots – the means to communicate with a website and get a response without using email. But in reality, these chat bots are glorified call centres with human beings providing the answers – the only difference is that you get your answers written down on your computer screen instead of spoken in a phone conversation with a call centre operator. While they can work well in some applications, for example when needing help with something technical, I find them really annoying when shopping on a retail online website – they just keep popping up when you’re trying to make a purchasing decision in the hope that they can sell you even more stuff you probably don’t want anyway.

Call Centres

If you’ve ever tried to talk to a human being in a call centre without incurring a lengthy delay waiting on the end of a phone, you’ll know how incredibly annoying and frustrating it is. The recorded message is rarely honest about how big the queue is and how long your wait might be, and you dare not leave the queue because chances are the wait will be even longer next time you try. And those annoying prerecorded messages that are repeated at 30 second intervals (in between the usually awful prerecorded music) telling you how important you are to the organisation – the same organisation who won’t even answer your calls - and suggesting, if you’d prefer, to go onto their website to make contact, are enough to push you over the edge.

I’m sure call centres were originally introduced to save time and money, usually resulting in the company laying off their trained and dedicated (and often long-standing and knowledgeable) employees to be replaced by casual or temporary workers who are faceless, nameless people sitting in offices akin to battery farms, reading from a script, and often in different countries far removed from the organisation they’re meant to know about. And there never seems to be enough of them to deal with the volume of calls being made, hence the ridiculously long delays in having your call answered.

I’m not sure who actually saves time and money, it’s definitely not the long-suffering consumers who eventually buy elsewhere if they can, and it’s unlikely the company who finds themselves ranked at the bottom of customer satisfaction surveys.

Search Engines

For many people seeking general information, the advent of the internet and the plethora of search engines was like manna from heaven. At last, a way to find the information you’re looking for without leaving your desk or sitting on the end of a phone for most of your life. Yay! Until Google Ads came along that is, and you now find yourself wading through screeds and screeds of ‘sponsored’ information which is just another name for paid advertising which has been dressed up to look like genuine search results. These days, the actual information you are seeking is often several pages away with a pile of data to read (or scan) before you get anywhere near what you’re looking for. Who has time for that?

There has to be a better way for organisations to impart useful information in a more user-friendly manner.

Semantic Searching (AI)

Using semantics to find information is the breakthrough we’ve all been looking for. Imagine going onto a website and being able to ask a question in almost any way and getting an answer that is specific to your question. Especially if your question is a bit technical or specific to a certain scenario.

For example, I recently completed building work to my house and wanted to apply for a Certificate of Code Compliance from my local council. I searched their website (both manually and using their search facility) for information about the application process but couldn’t find what I was looking for. My only option was to phone their office and the operator promised to get back to me, which she did some 20 minutes later. She told me that she’d found the information on the council’s website but it was buried deeply and she doubted I would have found it on my own. She gave me the information I was seeking and I duly completed the application form. This was a relatively painless experience because I was dealing with a small-town council that still employed their own staff. If they had a semantic search facility, I could have got this information really quickly on my own, as a bot would have done the searching using an algorithm.

Semantic searches rely on words that are refined to specific topics, and the information, which is already available, is collected and stored in vast knowledge bases in a format that responds to a range of questions asked of it. The good thing is, if the answer isn’t known, a search request is sent to the host organisation who will research the possible answers and these get loaded back into the knowledge base (as well as emailed to the original enquirer) for future enquiries. So, the knowledge base keeps growing and refining every time it is used.

It’s any body’s guess how long it will be before all information searches are conducted in this very efficient way, but given the leaps and bounds since the start of the Information Age began, I’m hoping it won’t be long at all. Bring it on, I say.

What can go wrong?

The biggest thing that worries me about the use of AI when searching for information is the trickery or potential abuse – in many cases we’re being constantly badgered to buy more and our profiles have been closely manipulated by algorithms to learn our inner-most shopping habits. However, it could be much more sinister than that. Some of the world’s biggest players have concerns about reigning in the ‘beast’, should it get too clever for our own good, and have urged AI labs to stop training systems to surpass GPT-4. They’ve also asked for a six-month pause on the advanced development of GPT-4 (Open AI’s latest large language model).

In the words of Brad Smith, president and vice-chairman of Microsoft, AI has, like any other technology, the “potential to become both a tool and a weapon”. He goes on to say “that is why we need human control to slow things down or turn things off.” An article in CNBC, where he gives an exclusive interview, goes on to point out that Tech leaders around the world have warned that AI poses a human extinction risk on par with nuclear war after generative AI-powered chatbot ChatGPT exploded in popularity for its ability to produce humanlike responses. Read the full article here.

Smith is quoted as saying, in regards the ChatGPT, “It is a tool that can help people think smarter and faster. The biggest mistake people could make is to think that this is a tool that will enable people to stop thinking.”.

I think the concern is not about people ‘not thinking’, (social media has already taken care of that for a lot of people) but using AI for evil means. Seeing the growth and sophistication in scamming in recent times, and knowing that AI in the wrong hands could lead to a new level of misery through theft, really worries me. It was largely thanks to the poorly-written, typo-ridden begging letters that alerted us to scams in the past – how will we spot them when they are written in perfect English, capable of passing the closest scrutiny?

A recent post by a member of the LinkedIn community commented that it was no longer necessary to allow two spaces after a full-stop (period in the US) when typing a sentence – they observed that this was something that people over the age of 37 did and was a remnant of the old days of using a manual typewriter and was no longer necessary. My interpretation of their post was that it was more likely a warning to those older than 37 to align their formatting with AI so that no-one knows that Chat GPT actually wrote the content. But then I might just be an old cynic…

My late father, who died in 1958, long before technology had given us television sets in our NZ front rooms, apparently used a saying that went “believe nothing you hear and only half of what you see” (original source was either Benjamin Franklin or Edgar Allan Poe). As a child growing up, I always wondered what this meant. As I got older, I realised he was onto something. Now I think it should be “believe nothing you hear and nothing you see”. Oh dear.