The GUIey Middle of Artificial Intelligence

The basic premise of artificial intelligence, to use enormous amounts of data to find out new things, is easy to grasp. If any one of us had the time and stamina to study a million photos or stories about a thing, I’m sure we’d come up with insights about it too. 

Business products emerging from current applications of artificial intelligence are also logical and simple to get your head around. Smart thermostats sell because they are convenient and deliver energy savings. Marketing approaches that analyze shopping patterns to suggest items people are likely to buy are winners in retail for their potential to increase sales.

How does AI get from data analysis to creating desirable products? In diagram version, this seems to me:

A few hypothetical1examples:

1. Using AI to improve diagnosis of medical images. Input: One hundred thousand pathology slides of renal cancer and one hundred thousand slides of normal kidney tissue. Outcome: Improved differentiation between normal and malignant kidney biopsies. Doctors win because the accuracy of diagnosis increases, saving healthcare costs by prescribing the right treatment for patients. Patients win because they are either can carry on their lives disease-free or have greater certainty in the treatment they need.

Mysterious GUIey2inside: What is the AI looking at to distinguish between normal and cancerous cells in pathology slides?

2. Using AI to improve traffic flow. Input: Every car in the city communicates its starting point, destination, and real time location to a central database. The goal is to send a uniform volume of traffic via every available route so that none are over-used or under-used. The outcome is a clear win – optimum travel efficiency for everyone, saving time, auto costs and impact to the environment by decreasing energy consumption.

Mysterious inside: What is AI doing to manage all the permutations and combinations to direct even traffic flow?

The two examples are different. In the first one, the criteria AI uses to distinguish between normal and malignant cells are the mystery. Pathologists could list the traits they use to make a decision when looking down a microscope, but is AI using the same ones? In the second, it’s the speed and capacity to deal with volumes of users that’s amazing. It’s not difficult to suggest the best route for your mother to take home, based on knowledge of traffic patterns at the time of day in your home town, but who could do that for 3 million occupants of a city simultaneously?

I’ve read that we are unlikely to be able to extract the GUIey middle3from AI supported processes, due to the iterative nature of the learning. When a person really understands what they are doing, they can explain it. If a chef tells you their sumptuous meal resulted from ‘a little of this, a little of that’, they likely know exactly what went into the dish, but aren’t telling to protect their trade secrets. If my mechanic tells me they are basing the diagnosis of what’s wrong with my car on some data from other cars but doesn’t know which models or what kind of data, I’m looking for another mechanic.

Is not knowing how AI works any different than not knowing the detailed working of automobiles, or any other complex object or process in modern life – elevators, mortgage documents, dental implants? The fundamentals of the car I get – the energy of exploding fossil fuel is converted into angular momentum that torques the axels and moves me, in my steel and plastic carriage, to where I want to go. The business model is also easy – the speed and convenience of reaching destinations in relative comfort with the added efficiency of carting a group of people, sheets of drywall, or my dogs with me. There is someone who can explain ABS brakes, how the muffler is connected to the engine, and all the other components that make a car function. With AI, either by design or trade secret, the explanation is hidden.

We need to know the mysterious processes that AI systems use to derive new knowledge from the volumes of data consumed. Forget proprietary algorithms. This is brave new territory we are entering and transparency is important so we can be sure we are operating safely and ethically.4

History is full of examples of embracing new things without a full understanding of the implications5. From that, a machine would learn that we need to know how things work before we can use them safely.

——

1Both of my examples are likely to be real enterprises but staying hypothetical is better for this discussion.

2This is a pun on GUI type computer interfaces, which use icons, rather than typed commands, to tell computers what to do. GUIs make programming simpler. I’m suggesting by making things simpler with AI, we are making them less transparent, dissectable or amendable to understanding how the parts work together to create the whole. Less concrete. More gooey. Gooey-er. Soft and flowing, changing shape easily.

3I do know that the process AI uses is a very large series of logic functions, of the sort: if X does Y, then A is the outcome. If X, K and J, do B, then L is likely to happen. If X does Y but K does something else, and it’s Tuesday, then Blue is the right answer. Etc. Oh, and the AI may start with a bunch of logic statements but change them on the fly as more data comes in or if in testing a hypothesis, it doesn’t deliver satisfactory answers.

4For many examples, read ‘Weapons of Math Destruction’ by Cathy O’Neil

5A few examples that spring to mind – nuclear weapons, cigarettes, social media, plastic, many types of home insulation, lead paint, breeding of dogs, trans-fats, mortgage backed securities.

Please follow and like us:
error

Ask What AI Could Do for You

Embrace new technology. Change life.

Who wouldn’t want that? Me, sometimes.

AI is the next big thing in disruptive creation. Errr, creative destruction. It’s not so new. Forms of AI have been embedded in commonly used products for decades: auto-correct typing, suggested products you may like, search results. Now, it’s becoming ubiquitous. The projected capacity of AI to make life easier is celebrated in investor conference calls, with services such as arranging for transportation to the airport when the AI knows you’ve bought a plane ticket, or ordering another box of laundry detergent when the one you have is about to run out. 

This makes me ask (and you can too), what would I love AI to do for me? Would it be great if AI ordered all my household consumables? Not really. I’m proud of my system for keeping a sufficient supply of life’s necessities (food, drugs, cleaning and pet supplies) on hand. It isn’t a big deal. If it is for you, or you just hate doing it, ok, I’d invest in that AI, if there were enough of a market to justify it.

Great new business models remove the pain of a current task, solve an existing problem. What do I see as really annoying, inefficient situations I would pay handsomely to change?

Here’s a starting1list:

  • the awkwardness of software updates – stop making me have to stop and think about something I’ve learned to do intuitively, like find the weather app on my phone screen.
  • the uncertainty of hiring competent contractors, plumbers, landscapers, auto mechanics. 
  • knowing when something I do regularly is going to change and how my life should adapt. When the bus schedule changes, I want to know if I need to get out of bed earlier, not just that the schedule has changed.
  • gardening solutions. Random bugs eat my leaves and buds. Critters steal my veg. Anticipating this, as preventative measures are likely the most effective, would be awesome.
  • interpreting what my cat says, translating to english. Seriously, why don’t we really know what ‘meow’ means, after domesticating cats thousands of years ago?

I happened on an application for AI that I didn’t know I needed until I needed it in a hurry. It required getting information from a series of government and corporate entities, late on a Friday afternoon, before a long weekend. And I got it. Because it was information that each entity stored electronically. So emails were generated to use the info to answer my questions. In 10 minutes! Huzzah!

There are probably many more services I consume irregularly that AI could speed up. From what I’ve read, the sorts of process AI is expected to be used in first are industrial/business applications. This means that many of the best uses of AI won’t be noticable to us consumers except in declining prices, faster delivery or a better selection of options.

Why my cautious approach to AI? There are many AI applications that I imagine would take the fun out of life. Anything that requires creativity. Or some combination of serendipity and knowledge. Interior decorating. Discovering new restaurants, clothing lines, bands, books to read. The whole point to discovery is that it’s random. If something tells you where to find it, that’s ok if all you wanted was to get the thingy asap. Roofing shingles, a new muffler, parts for your appliances, or shoe laces for your winter boots are like that. For other items, there’s the thrill of the hunt, randomly happening on the perfect wastebasket for the downstairs bathroom, shoes to go with your suit, or a gift for your three year old.

I strive to challenging myself to achieve more, learn more, do more, in physical, intellectual, and economic realms. If AI made it all easier, I’d cease to grow, learn or improve. Proponents of AI might say the technology would allow me to stop wasting my time on parts of life that don’t challenge, so focus is on improving in important areas. AI might even lead me to the next, more enriching challenge.

What do I wish AI would do for me? Take care of the annoying things and leave me the interesting ones. Bearing in mind that what I find annoying, you may find interesting, the key is to make everything more efficient but make the high efficiency version elective. A mundane example of this is that grocery stores sell loaves of bread, but also all the ingredients to make bread from scratch.

That’s real intelligence, delivering what each customer wants.

——

1I’m writing as an individual consumer. I may be in a demographic of one, which doesn’t make for a good business model, unless the product costs millions of dollars, which I don’t have, so forget that. However, more than likely I am in a demographic of significantly more than one, as most of us are.

Please follow and like us:
error

AI Personal Assistants – The Death of Shopping as we Know it

Predictions are, in the near future, we will each have a personal assistant with artificial intelligence (AI)1 that runs our life. It’ll order household items before we run out, book social engagements, reminds us of upcoming events and related purchases (like birthday gifts, a bottle of wine for the hostess, or a new outfit to wear to the party).

More elaborate predictions have the AI constantly searching for better deals on services like vehicle sharing, archery lessons or landscaping services. It’ll sample the news wire for updates on unhealthy foods or ethically produced music, keep up to date with product reviews (posted by other people’s AI personal assistants) and use this collected wisdom to amend our purchase decisions (which the AI made in the first place, so we won’t even know).

This got me to imagining the end of marketing as we know it. No more emotional buying decisions. Every single purchase would be made with the maximum amount of data and, hopefully, solid facts.

Why would an AI be interested in brand loyalty? An AI would access all available information to determine if the latest version of a brand name item delivered on the quality expected, and if not, find another brand that did. Far fewer buying decisions would be based on the logic ‘I’m buying Apple because Apple makes good technology’. Your AI would buy Apple if there was proof it was the best available technology. And the proof would come from objective tests and the unbiased reports of AI’s everywhere (because why would an AI lie?).

Trickier is image, prestige, lifestyle or that thing where you buy a certain brand because it reflects who you want to be. Would your AI get that, have the same image of you as you do? That you wear a certain type of sneaker because people who share your values do.

Then there’s the ability to forget things you prefer to forget. Like booking a dentist appointment because you don’t like going to the dentist, so putting it off another month would be fine. Would your handy personal assistant let you do that? The dentist would be happy if you came back more often, so the dentist’s AI would encourage yours to book, maybe offer a discount. The same rationale could apply for the vet, furnace cleaning, arranging a visit to those relatives you find tedious, getting the oil changed in the car you jointly own, and a few dozen other things that fall into the category of adulting ( willingly doing things you know are good for you but are unpleasant, no fun, boring etc).

Then there’s retail therapy. Could your AI pick out the perfect new sweater for you, when you don’t need a new sweater and can’t afford it, but accidentally yelled at your boss, spilled milk on your toddler, and got a ticket for not going through a green light all in one day?

Is having an excuse to get out of the house a thing any more? Shopping used to be a good neutral destination that always worked if you needed something to do or to get away from the humans you lived with. You can’t get your AI to do that for you. Unless it pretends to be your friend who has to meet you at the mall.2

There will always be new ways of doing things. But humans are humans. We learned to live much of our life online, but we shop for more reasons than to get stuff. We also forget things on purpose. We act on our emotions because that’s what makes us human.

I think I’ll sneak out of the house, tell my AI personal assistant I’m on my way to the dentist, then cancel the appointment so I can go shop for stuff I don’t need, but want.

——-

1Purchased from a large tech company and embodied as a hockey puck-size matt silver thing that sits on the kitchen counter.

2If this sentence doesn’t make sense to you, please review a TV show or movie from the 1970’s for context.

Please follow and like us:
error

How Smart is Artificial Intelligence?

Cyberspace isn’t much like space at all. It’s crammed full of bytes of information, churning and frothing with intelligence agents who gnash and dissect the data in search of new knowledge, or at least something else to sell us. This is big data and at least one embodiment of artificial intelligence.

Recently, I heard an elegant explanation¹ of machine learning, or the ability of machines to create programs and algorithms that deduce things that they haven’t been programmed to – how machines learn. Consider what would be involved if you had to write program to tell a computer how to distinguish between a cat and a dog. I’d put together a logic chart and add up the cat vs dog points:

Cat

Dog

Meow noise

1

0

‘ark-ark’ noise

0

1

retractable claws

1

0

floppy ears

0

1

stripes

1

0

lolling tongue

0

1

it’s ignoring you

10

0

it thinks you are the smartest, most desirable person in the world

0

10

thrax ignoringand I’m sure you can come up with many other criteria, some less than absolute, such as curly fur (much more common in dogs but not impossible in cats).

In machine learning, you’d give the computer a million videos labelled cat and a million videos labelled dog to watch and let it figure out its own algorithm to tell the difference. Who wouldn’t want the job of watching a million cat and dog videos? Most of us already have. I am curious about the computer’s algorithm: does it use tail wagging frequency, that silly whining noise dogs make, or hissing, as selection criteria?

What if after all that the AI comes to the wrong conclusion. It might decide the true difference between cats and dogs is that cats are the overlords of the planet and dogs are service animals. It’s easy enough for a mere human to decide if the computer has done a good job of differentiating between the two animals. But what happens when they start predicting things we have no prior knowledge of, like how long a pair of socks will last?

And this is a trivial application of artificial intelligence. There is so much data out there, silly names for bigger numbers have emerged. According to this BBC article, 2.5 exabytes (billion gigabytes) of data were generated in one day in 2012 and the US National Security Agency has the capacity to store a yottabyte (one thousand trillion gigabytes) of data. That’s a lot of Facebook likes, tweets, diagnostics at the auto-mechanic, GPS locations, term marks and everything else. If we set AIs to learning from all this data, it seems like a tremendous wealth of knowledge will emerge. This might fall in a few categories:

1. Important and life saving intelligence such as diagnosing serious health events like heart failure and intercepting terrorist plans, so interventions can be made earlier.

2. Efficient systems, such as automated traffic flow to relieve congestion or business processes like finding items (books, events to attend, cheese) people might be interested in based on their preferences.

3. Predictions – varying from novelty (suggestion of what the name of you next pet should be) to kinda useful (prediction of what your partner might like for dinner tonight) to downright world changing (motivational media reports – this is one of my personal dreams).

The biggest question in my mind right now is how do we know if the machines are right?

Sure, we can test each conclusion the machine reaches after it’s made but that will take some time, especially if it’s a long range projection. And who owns the predictions? Is information about me that I don’t know, like what diseases I will develop in my old age, my personal information?

Yikes, I don’t want to go a bad place with such a potentially good thing. Like most new technologies, there is the possibility of misuse and misinformation with machine learning and artificial intelligence. Maybe we can use machine learning to figure out how to avoid the misappropriation of information for improper purposes. That would be cool. A truly self-regulating system.

—–

¹ I believe it was from Steve Brown,Chief Futurist and Evangelist, Intel at the Plenary Session ‘Innovation: Steering Disruption’ of the International Economic Forum of the Americas in Toronto July 8, 2015

Please follow and like us:
error