We interact with AI almost daily — for example, calling customer service and conversing with a disembodied computer for five minutes before being able to talk to a human being (if at all). And how about after ordering laundry detergent online, getting six different pop-up ads for all kinds of cleaning products.
Those are some AI applications that annoy me the most. Although, I don’t mind YouTube bringing up one video after another, helping me discover new things I might enjoy.
Will AI Take Over the World?
And then there’s the debate over whether AI will overtake human intelligence (HI?) — see the hoopla about ChatGPT. I personally think that’s a silly debate. There’s no way, at least in this century — if ever — that a computer will be able to think like a human.
AI can be a useful adjunct to a human brain, but even with ChatGPT, humans have to keeping correcting and teaching the software.
AI doesn’t think the way we humans do. We can make associations among all sorts of different sensory inputs, memories, and feelings, and come up with creative ideas. We can discuss those ideas with others and decide whether there are things we would like to change. Our thoughts might be used for practical purposes, or sometimes just to give ourselves pleasure, or ways to cope with pain.
What I Like About AI
AI machines are really good at digesting large amounts of data to produce useful information. But, ultimately, it’s we humans who have to decide which of that produced information is meaningful and what, if anything, should be done with it.
The Tech Briefs website offers many insights into how AI is being used in ways that I think are extremely important.
Sensor-Based Network Applications
AI is used in a variety of applications to enable sensors to be “smart.” It can help meet three important challenges for sensor-based networks:
- Low latency
- Low power
- High security
Low latency, the amount of time required to process raw data into usable information, is especially important if a quick data-based decision has to be made in an operational network. For example, in an Advanced Driver Assistance System (ADAS), there’s little time to waste in making a safety-critical driving decision.
ADAS relies on combining inputs from different sensing systems like LiDAR, cameras, and radar — a process called sensor fusion. AI algorithms enable software to be used to optimize the way the data is processed so the sensor fusion is as fast as possible. (See: Artificial Intelligence in Cars — Inside the Brains).
Low latency is also critical for factory automation systems, where manufacturing processes are controlled by data fed back from machines in real time. By embedding AI processors in sensors, less data has to be sent to a central processor, or even more critically, to the cloud. This reduces the amount of data that has to be moved — the more data that has to be moved the longer it takes.
Data analytics performed with smart edge devices can be done at milliwatt levels, compared to the much higher power needed for calculations at data centers. AI can also be used to decide when it is critical to send data to a processor or to otherwise allow a sensor to run in a low-power standby mode.
If analytics are performed by AI-embedded sensors, a smaller amount of sensitive data needs to be sent out to the cloud, where it is more vulnerable to being hacked.
Just as with automotive or factory applications, there is a need to convolve different data points from a variety of sensors to gain important health information.
“Artificial Intelligence Tech Set to Transform Heart Imaging” describes an AI application that can produce better imaging of scar tissue in someone’s heart than the traditional approach, which requires injecting a contrast agent.
The article “Artificial Intelligence for Astronauts Monitors Patients at Home” describes how data systems used to monitor astronauts during space walks can be adapted to Earth-based uses. California-based Ejenta has licensed the NASA software and uses it to collect data from commercial health and fitness monitoring devices to create individualized reports about a person’s overall health. It can then report this to the person themselves as well as their healthcare team.
Another article describes how AI can be used for more than diagnostics to make people’s lives easier. AI software is being used by researchers to enable exoskeleton users to “…climb stairs, avoid obstacles, or take other appropriate actions based on analysis of the user’s current movement and the upcoming terrain,” eliminating the need for the user to analyze the situation and decide what actions to take.
The Way I Feel About AI
While these technical applications might not make for fun chatter on social media, they represent some of what I view as game-changing technology to improve our lives in important ways.