Mark Conde
 |
It is hard to decide if AI influence on our lives is good or bad unless we strive to understand what it is and how it is used. As I indicated in the last column let us look at examples from four types of AI: media, context, sensors and predictors to examine the good and the bad/ugly.
There is so much information on the internet that is often hard to digest and focus on what is important to you. You gather information with a Google search but must wander through many entries extracting what makes sense and form one view on a topic.
Using a context AI tool like ChatGTP does this all for you looking at relevant information, all at once aggregating an answer that focuses on what you want to know. This is a good example where AI becomes an enabler for humans and take us to a higher level of information use.
This human enablement is a capability we would like our doctors to have. They must train for years to know how to diagnose a health problem, and it is growing more complex every day.
AI media and context processing can be their partner to look at a set of symptoms and suggest a possible affliction and how to treat it quickly. Processing radiology pictures though AI media processing has proven to be very successful finding anomalies the human didn’t see.
I would want my doctor to have the capability of many doctors at once, all looking at my issue to find the best solution quickly.
Will this replace my doctor? Not likely, because we still need the creative capability of humans to think about the things that are not seen. That same AI helper may never have been trained to ask questions about my behaviors like do I smoke thus missing a critical part of my health problem.
Research is seeing huge ways to use all types of AI. A cancer data base in Georgia, GCCS, uses context and media AI to interpret pathology, radiology and oncology reports to find patterns that are discovering more effective treatments.
Public health departments use AI predictors looking across historical data to find what specific flu viruses will show up this year, so they know how to load our vaccinations.
Have you seen the cool new AI media tools that can correct fuzzy grandkid pictures with a swipe across your phone? I bet you appreciated that your new car automatically hit the brakes when that car pulled out in front of you using AI sensors.
I could keep talking about the many good uses, but we need to explore the bad and ugly, too. It is critical that we understand the bad things you hear about and how we will protect against them. The next column will dive deeper into this and talk about how we protect against the evil uses and bad AI development.
One of the bad things is the ability to create fake and very realistic videos and voices to trick you. One AI-generated scheme may be a phone call from a relative who sounds exactly like them, and they need money for an emergency. Protect yourself by setting up a
“safe word” that you ask
for to confirm it is really them.
Other bad areas we will explore in the next column are job replacement, incorrect and biased AI training, and using information that is private or copyrighted.
I want to make sure you understand these things and how AI can continue to be an enabler for humans into the future.
Mark Conde of Murphy is a retired chief information officer and self-described technology nerd. Email him at
jmconde818@gmail.com.