 |
| Machine Learning. |
According to Wikipedia, Artificial intelligence (AI) is intelligence exhibited by machines. In computer science, an ideal "intelligent" machine is a flexible rational agent that perceives its environment and takes actions that maximize its chance of success at some goal.
The term “artificial intelligence” sometimes conjures up images of the distant future, and a time when self-aware robots take over the planet and make human capabilities irrelevant, because the computer overlords are way better at everything humans do. While that vision may be workable for a hair-raising Halloween movie script, it’s way off base when it comes to the realities of the current generation of artificial intelligence (AI) - CIO
If you tackle this story from the highest level, AI isn’t currently a force that is poised to overrun humans. Instead, AI is simply a group of technologies that will increasingly be used to augment human capabilities, and make us better at the things we do best. What’s more, AI isn’t a story set in the distant future. It’s here today, and improving our lives in countless ways.
Of course, it would be naïve to be too Pollyanna-ish about the risks of AI: We have seen throughout history that powerful technology can bring both benefit and harm. As a technologist, my goal is to keep a watchful eye on the risks, while focusing on advancing the life-improving benefits that AI systems can deliver.
When it comes to augmenting human intelligence, there are some things that AI systems are particularly good at. One of them is processing at scale and another is continually learning from experiences—two machine-driven capabilities that extend the limits of human capabilities.
Let’s take a simple example that we all know well. When we use a search engine like Google, we’re experiencing the power of AI to augment human capabilities. The human brain isn’t capable of searching through billions of web pages to find the exact information we are looking for, but Google does this every second of every day. That’s processing at scale.
And that’s only part of the story. As Google responds to search requests, it continually learns from its experiences and gets better at understanding the information users are looking for when they search on certain terms and phrases. When the top-ranked suggestions are the ones the users click on first, Google knows it got it right. When that’s not the case, and the user continues the search from difference angles, the algorithm learns from the experience, so it can do a better job for users in future searches.
A backstory here is that AI is purely data driven. In simple terms, if an AI algorithm doesn’t have data, it can’t be smart. Computer systems are powerful only to the extent of the data that they have been given. Humans set the goals for computers, and then make sure the computers get the data they need to both answer and ask questions.
While AI systems are here today, working for us in various ways, in the years to come they will be much more powerful and will be much more present in our lives. When I say that, I’m not looking into the far-off future. I’m focusing on a few years down the road.
To take an example from the medical industry, the capabilities of physicians will soon be augmented by AI systems that search through massive amounts of data from various sources to isolate information that is relevant to a particular patient’s condition right now—as the patient is being examined by the doctor. If a patient is diabetic, the system will show the doctor everything that is relevant to diabetes, all the way down to long-term care things like past foot exams. This is the promise and opportunity of precision medicine.
 |
| Sponsored Ad |
Physicians, of course, are very good at understanding what is relevant and what isn’t relevant in a patient’s chart. But they aren’t capable of searching thorough gigabytes or even terabytes of data to identify relevant information that will help them understand the true state of the patient, or what is changing right now. While AI systems aren’t there just yet, they will soon be quite good at this task.
Let’s take another example of AI in healthcare, simply because it is a field I have focused on often in the past. Ultrasound machines give healthcare professionals a view into the bodies of patients, all the way down to close-ups of individual organs. Historically, these machines were used only in hospitals and only by technicians with special training. In a breakthrough that is coming our way, we will soon have portable ultrasound machines that can be used outside the hospital—for example, by emergency services personnel. But there’s one big catch here: The personnel who use the machines outside the hospital aren’t likely to have the extensive training of the ultrasound technicians who work inside the hospital.
To get around this problem, device manufacturers are working to build intelligence into portable ultrasound machines, so the usefulness of the device is not tied to the skillset of the user. The machines will have the intelligence to understand the orientation of the patient, make decisions about where to scan and understand what they are scanning. With this ingrained expertise, portable ultra-sound machines will augment the limited ability of humans, and then pass the results on to humans to be interpreted and used to guide clinical actions.
As those examples illustrate, artificial intelligence is going to help us get better at the things we do best. It’s never going to replace the need for human intelligence. It’s only going to augment our capabilities, and extend our reach in places where we are limited.
While it won’t form the basis for an entertaining Halloween movie about machines gone wild, artificial intelligence will enrich our lives in countless other ways. So stay tuned for upcoming episodes.
0 comments:
Post a Comment