Skip to content

Hallucinating the AI of 2033. (part 2)

Previously, I looked at the disruptive forces of the Internet and mobile phones to see how they changed life in a decade. The goal was to help me hallucinate1 how AI will impact our lives ten years from now. 

It turned out that the ways that the Internet and mobile phones changed life were only so helpful to predict how life would be changed in 2033 by AI. Surprise – it’s difficult to predict unexpected things. Nevertheless, I have a lot of ideas about what might be changed ten years hence because of AI. Some are fanciful, some directed towards North American consumer inconveniences, others humanitarian. 

My predictions used two characteristics of AI as we currently know it – its ability to analyze large amounts of data and the language generative skills that are all the current rage. 

In the categories previously observed:

1. It didn’t happen overnight: Gradually, more and more parts of life will involve bits done by an AI. 

After the initial couple of years of wonderment, we’ll nod our heads as each everyday process incorporates AI, making it more efficient, for someone2. Five years hence, we’ll start wondering when AI is going to fix the remaining slow processes, and be pulling out our hair in year ten when the last few archaic systems are clunky because they don’t use AI (like wondering why your doctor still insists on FAXing other doctors today).

Before 2023, AI had its hands (subroutines? input functions?) in a number of processes. For example:

  • keeping track of when the house is cold or hot to make energy-saving adjustments, 
  • writing standard responses to customer queries to let them know a human will be in touch soon, and 
  • analyzing cause and effect to customize insurance rates, loan approvals, medical diagnoses, grocery lists.

Now, it seems everyone is trying out natural language processing AI’s (like chatGPT) to do anything and everything that requires scripting an answer, including: 

  • students doing assignments, 
  • companies answering customers, 
  • employees supplying their bosses with information, 
  • writers getting prompts, 
  • people being creative. 

This experimental phase will die out once clearly useful applications of natural language processing are established. 

A key point: I expect AI to fit into everything we already do, rather than make new things. It’s primarily a tool, multi-purpose, like WD-40, to lubricate all the sticky situations in life.

I’d like natural language processing to make truly helpful automated attendants to answer the phone at every large corporation and government. 

This is the opposite of using language processing to answer; it is using language processing to listen. 

Right now, automated voice attendants have a limited repertoire in their responses. Imagine calling a company with an situation you know is complex and needs a non-routine solution, and having that immediately understood and getting right to solving the problem. 

The reason often given for why it is so difficult to wade through company contact centres for an answer to your real problem is that so many people call with trivial questions. AI should be deployable to answer an infinite number of trivial questions, pleasantly and accurately. AI could repeat the return policy is on a specific order even if a million customers are simultaneously asking the same question, without getting bored or pissed off. 

The AI attendant could be better at reasoning with difficult customers, for business benefit. There are just some completely unreasonable expectations. A porcelain cup is dropped on the floor. It breaks. Might the AI calmly explain the inability of the material to resist sudden impact is a characteristic of the product and hence its appeal. The AI wouldn’t fear being fired for annoying customers. And the AI would resist saying anything to the customer to keep them, because that would be illogical.

Perhaps we will get to the point of truly personalized customer service which doesn’t lie to you, because business ethics dictates that customers shouldn’t be lied to, even if it is to their benefit. 

Other applications: Recycling bins will express their approval or disapproval for what can be placed in them. Better yet, all containers will auto-sort so they end up in the right recycling process. Then, citizens can go back to the bad old days of putting everything in one (not)bag, and waste is sorted post-collection to where it needs to to be environmentally sustainable.

In the future, AI could serve our entertainment needs by whipping up a movie for us based on our wishes: favourite actor, some vague plot (e.g. slaying dragons, work promotions and cheese addiction) and favourite genre (e.g. documentary). 

After the novelty wears off, I suspect on-demand creative works will have limited following, because: 

  • humans want a viewing experience they can share with their friends (why else would a zillion people want to simultaneously want to watch the last episode of Game of Thrones) and 
  • the joy of movie going is to see what someone else’s imagination can come up with – many someones, including the writer, director, producer and all the actors. 

AI might assist in home decorating, traffic regulation, drafting contracts, planning vacations, medical diagnoses, and many other activities that conform to rules, even quickly changing rules (periwinkle blue is dead, bring on the aquatic blue), or a complex tangle of them. 

2. Didn’t know I needed it but it sure is useful: … part 3

1 If AI can hallucinate about current reality, I can hallucinate about the future of AI.

2 There are things made more efficient for the offering company by technology that are less efficient for their customers, or employees. As a consumer, most online forms spring to mind.

Thanks for reading.

If you'd like more,

sign up to receive awesome content in your inbox, every week-ish.

Signing up is only for updates when new blog posts are added to my site. If you want marketing spam, you'll have to look elsewhere, but, really, anywhere will likely do 😎

Share this post, if you like...

Leave a Reply

Your email address will not be published. Required fields are marked *