I’ve sent out three Daily Lessons recently with various AI use cases.
- Products with Personalities? My Interview with JD About Conversational AI
- 3 Ways AI Is Transforming Fashion: My Interview with JD Vice President of Cloud & AI (Jeff’s Asia Tech Class – Podcast 33)
- 3 Lessons in China AI/ML from Artefact (Data Consultants and Digital Marketers)
And I think use cases are a good way for executives to approach the subject of AI. Just hunt industry by industry and see what companies are using AI for and not. And what is working and not. And make sure you look at industries other than your own. I find it a good way to stay on top of the subject, without getting into software coding and such.
A bit of review.
The Four Waves of Artificial Intelligence
In a previous update, I used the four AI waves as described by China AI guru Kai Fu Lee (author of AI Superpowers). Here is a slide he presented at a China conference.
Basically, he argued that the first wave was Internet AI.
This was online companies started applying data analytics and AI to information they were already automatically gathering and processing. It was a natural move for software-based companies to use the latest types of software on their data and processes. Think companies like Google, Amazon, NetEase, Baidu, Facebook, Taobao, WeChat, JD, Meituan, and Toutiao.
Wave two was Business AI. This was when real world businesses started to apply AI to their well-developed systems for data and decision making. Like with wave one, these companies had long been collecting information about their business. So this was a lot about mining this data for hidden insights. And while businesses are usually quite good at making correlations based on strong features, AI can be very useful for finding correlations based on weak features. Think of companies like Palantir and IBM Watson.
Wave three is Perception AI. And this is about digitizing the physical world. It requires sensors, microphones, cameras, IoT, new highways, and so on. And the physical world is being turned into data that AI can start to run algorithms on. I think this wave is pretty amazing. Because up until this point, machines really only knew what we told them. We had to type in or load information for them to know much. But it turns out computers have really good vision. And they can increasingly see and process the physical world for themselves. They can gather data and process it, without our involvement. Think products like Face++ and Amazon Echo.
Finally, there is the wave four, which is Autonomous AI. This is when AI prediction starts to be joined by autonomous decision-making. As mentioned, you can think of AI as “cheap prediction”. Well, autonomous AI is when this cheap prediction is joined with decision-making and the cars start to drive themselves. So that is companies like Tesla and Waymo. And the whole frontier of “AI meets robotics”.
Ways Machines Are Better Than Humans
Machines are a lot bigger, stronger and sturdier than humans. It’s not even close. Large airplanes can fly +3oo miles per hour. Tractors can demolish buildings. Robots can go to the Antarctica and to the depths of the ocean. Satellites can go into deep space. Machines are physically able to do things way beyond humans.
Well, the same phenomenon is happening with cognitive abilities. Humans cannot compete with AI in memory or processing speed. Computers have perfect memories. They can solve equations and calculations at lightning speed. And they have tremendous precision and consistency. A computer can do a calculation or retrieve a random fact a million times without mistake. Without ever getting tired. Forever.
We can’t do anything like that. So it’s worth keeping in mind that when it comes to memory and processing speed, we are outgunned. Don’t try to compete with machines in these areas.
And now to the point of the update.
Ways Humans Are Better Than Machines. And How AI Can Make You Really Dumb.
There are things that humans are much better at in terms of intelligence. Or, to put it another way, there are things that AI is just terrible at. There are tons of situations where AI will give you meaningless and wrong predictions.
Here’s a big one:
AI Can’t Do Analogies.
Douglas Hofstadter (Nobel prize for cognitive science and how our brains work) called analogy “the fuel and fire of thinking.” When we see an activity, read a passage, or hear a conversation, we are able to focus on the most salient features, the “skeletal essence.” And we are able to extrapolate to other situations. Looking for both similarities and differences.
“True intelligence” is the ability to recognize and assess this type of essence of a situation. And we are really good at it.
In particular, we are good at doing this with analogies. We collect and categorize human experiences. And then we compare, contraste and combine. We are great at analogies and determining the “skeletal essence”.
AI can’t do analogies. In fact, AI can’t think at all.
If you see a Picasso line drawing of a dog, you know it is a dog just by the outline. AI has a very hard time doing that because it doesn’t know what a dog is. It doesn’t know who Picasso is. It has no intelligence or understanding of anything. It just scans for pixels. AI operates the same way New Zealander Nigel Richards wins victory after victory in French Scrabble, even though he doesn’t speak French. He just memorized French words and then puts together combinations of letters. He can do it even though he has no idea what the words mean.
This is how AI functions. And that is why it can’t do analogies.
AI Can’t Recognize Bad or Fake Data.
Garbage in, garbage out is a real problem for AI.
First of all, AI is hunting for correlations but has no understanding. So if the data entered is bad or fake, the AI is completely unaware of this. Just as it is unaware of everything. A researcher studying a question and testing for a causal mechanism is far more likely to spot bad data.
AI also requires big data. It doesn’t work well on small data sets. But the more data you use, the more likely you will have bad, fake or biased data included. So big data makes the bad data problem more likely.
Basically, AI cannot identify bad data. Corrupted data. Incomplete data. Fake data. Or biased data. It’s a big problem.
AI Can’t Recognize Bias.
A lot of this thinking is from Gary Smith, an economics professor at Pomona College and author of the AI Delusion. And the question Professor Smith focused a lot of attention on in his book is bias. That how you choose your samples for testing can introduce bias. Survivor bias. Self-selection bias. And many others. I put this in the same category as Bad or Fake Data.
Big Data is Full of Meaningless Patterns and Correlations.
If you randomly generate 1,000 numbers, you are going to see all sorts of patterns that strike you as meaningful in these numbers. Because correlations and patterns are everywhere. And we think patterns are unusual and therefore meaningful. But they usually aren’t.
Random data always has tons of patterns and correlations. With no underlying cause. It’s just random. Mediocre stock managers can have great runs for 5 years. Average baseball players can have long runs of success for 15 games. If you flip a coin ten times, there is a +40% chances of a streak of 4 or longer. Meaningless patterns are normal.
And it turns out computers are super-efficient at finding meaningless patterns in really big data. In fact, this is arguably their primary activity. The vast majority of correlations they identify when grinding through data are going to be meaningless.
AI Is About Correlation. But the Scientific Method Requires Causation.
When testing something, you should:
- Avoid data mining. You always want to have a cause, not just a correlation.
- So start with a theory.
- Then try to use small, accurate, understandable data to test it.
- Ideally, you test your theory going forward with an experimental and a control group.
- You then want to retest going forward with fresh data. It must be repeatable. With new data.
- And always ask. Does this make sense?
AI can’t really do much of that list. It mostly just fits the data. Which is easy with lots of variables and coefficients.
So the irony is artificial intelligence is not really intelligent at all. It just has the appearance of intelligence.
That’s it for today. cheers, -jeff
- 4 Strategy Lessons from the Spectacular Rise of Sea / Shopee / Garena (Asia Tech Strategy – Daily Update)
- Dingdong and 5 Questions for Assessing Specialty Ecommerce Companies (2 of 2) (Asia Tech Strategy – Daily Lesson / Update)
From the Concept Library, concepts for this article are:
- AI – Cheap Prediction
- Smile Marathon: Machine Learning
- Specialty Ecommerce
From the Company Library, companies for this article are: