I’ve been imagining what functions AI may play in 2033 based on observing the impacts of the internet and mobile phones. Previous hallucinations here.
In the category, ‘didn’t know I needed it but sure is useful.’ Difficult to predict, because … didn’t know I needed it. Nevertheless, a few wild conjectures:
Augmented human vision. See like a dog, helicopter, or cat (no wonder they’re so aloof).
Really good translation of foreign films. Voice-overs that not only change the words to another language but change the way the actors’ lips move so it looks realistic.
The cure for ‘there was something else I needed at the grocery store but can’t remember what’. AI in your vicinity will recall everything you mutter to yourself, such as ‘running out of celery’. And tell you when you’re at the relevant interface.
Communications from many retail outlets become more polite, realistic and welcome – the sort that come from the best salesperson ever. The one who contacts you just the right number of times, suggests only items that fit the time, budget, temperament and need. The one that immediately senses when you are not receptive. but knows the difference between totally not interested and interested later.
Those that don’t qualify never reach you because AI can engineer the best spam filter ever. This perfect spam filter removes not only generally accepted annoying messages but the stuff you personally find useless and annoying for your own reasons.
AI may sense your mood while you compose work emails, then select vegetables for dinner. Or correlate the weather, your shoes and a music play list.
It will integrate lots of stuff that hasn’t previously been connected. What will happen when the washing machine, stats on horror movies and a sudden passion for recipes involving gooseberries collide in your financial planning system: A new exercise routine? or Gift suggestions for coworkers?
If AI imitates your voice it could do all the nasty things you dread, like firing people, although this sounds unethical. AI could appear on video conference and make your visual image speak the words that convey those uncomfortable messages. This could extend to personal life, and conversations about the end of personal relationships. If the AI really could channel the right message, without getting emotionally derailed, it could be valuable. I’m sure there’s a viable business model.
Next category: Sure, it does all these things, but is it safe?
How much can we, as humans, let go of complex infrastructure systems and still be in control of our lives? I’m reminded of the great eastern seaboard (of North America) power failure of 2003 when millions of people were without power for many hours due to a software fault. In retrospect, we didn’t do too bad. The transit system failed, so we walked. Barbecued dinner in the absence of electric stoves. A bunch of babies were born 9 months later. We figured out what to do.
Humans are too resilient and resourceful to be utterly dependent on our engineered systems. We should design systems with this knowledge in mind. Questions will be asked until and beyond 2033 about how much control AI should have.
Meanwhile, looking into the crystal ball:
Many businesses pop-up that validate information produced by AI. Like grammarly, but for facts. This would be a good job for AI. AI to tell us if AI generated content is correct. There would only be a conflict of interest if all AIs were working together. Since each business would have its own, this wouldn’t be the case, except in a conspiracy theory science fiction movie, called something like ‘The AI Rebellion’.
This fact checking could extend to any human-provided information. Imagine a politician giving a campaign speech, with subtitles from AI that either validate or dispute the information. Or as commentary to ads.
Al could also be used to control the fraudulent use of AI generated creative works. It should.
Another Category: The Surprises, good and bad.
Most AI applications are components of commercial products, integrated into products by the manufacturer. I expect the good and the bad to relate to whether the implementation is more to the business’ or the customers’ benefit.
AI could be deployed to make it harder to provide deceptive answers by correlating the answer to other information, or information given in another situation, maybe 40 years ago. No more fabricated excuses about why you’re speeding, returning an item, or claiming expenses on your tax return.
We may get used to explaining ourselves to AI, as it will irksomely analyze things about you you don’t want it to, like your diet, heart rate and blood cholesterol, spending at various retail outlets and choice in partner (business, social and romantic). Something like two-factor authentication, only more invasive: “Are you really buying that chocolate cake for the kids?.”
Information everywhere. More than all info at the touch of a fingertip, information sprouting from every surface and interface our senses perceive. Being able to control when the information appears with a customized gesture could be useful. Current state of gestures to control information flow is a delicate repertoire, kind of like origami of the touch screen, but dictated by the touch screen manufacturer. Learning to filter the complex information both by avoiding the accidental gestures (like walking through a booby-trapped tomb) and limiting consumption sounds fabulously exhausting.
Librarians are super heroes. With so much potential for increased availability and overload of information, the relatively low profile career of librarian becomes the new ‘wanna be’ in 2033. Entrepreneurs replaced rock stars as the coolest people in North American mainstream media early this century. In ten years, librarian could be the envied profession, as they rein in rogue content spewing AIs, wrestling them into vessels of truth, or compassionately pulling the plug when the AI has gone too far.
Final Category: How will these technologies improve lives of more people around the world?
Will AI provide greater support of human needs for a great number of people? It should improve communication beyond current levels. Since the AI chatbot’s job is to process language, it may point out when two people are using the same word but mean different things, or a cultural different in interpretation, facilitating greater intercultural understanding.
Could AI help solve complex social problems, like homelessness or substance abuse? I hope so. Perhaps by identifying interventions for people at risk sooner, providing earlier detection of need and more readily available, easy to accept solutions. The analytical powers of AI may be able to see better ways to organize existing social programs so governments can direct their efforts more effectively.
This post was many hours in the making, most of which were spent herding my free-range imagination about what AI could do for, or to, us. These predictions are by no means exhaustive. I could go on, but won’t.
AI will have changed life significantly ten years from now, but it will take some thought to see how. Forty year olds will talk about how it was when search engines and wikis were all there was to provide answers to questions about life, the universe and everything.