SAN FRANCISCO, CALIFORNIA — When Prof. Amnon Shashua,president and CEO of Mobil Eye and senior vice president of Intel Corp., forecasted that language would be the new frontier in Artificial Intelligence in three to five years, I was intrigued.
We will have computer softwares that can read, comprehend, and analyze a text, as well as write a text. And it can read and comprehend millions of texts in a few seconds.
So it will now be possible for a computer to write a story by simply giving it a summary of your ideas. Voila! You have a new story. I suppose much better than what some writers or self-adulating writers can produce.
Just imagine its current implications to other aspects of our lives, be it in medicine, law, education, national security, economy, et cetera.
To date, an oncology cognitive computing system which uses AI algorithms has been used to generate treatment recommendations for cancer patients. I suppose much better than the best oncologist in the world.
After all, to update one’s self of the emerging or current state of medical knowledge, how many medical journals or medical researches can a doctor read, comprehend, and digest, let’s say, in a few seconds?
A robot judge is now being developed not only to catch and recognize unpredictable speech patterns, a rise in body temperature, as well as hand and eye movements of defendants or witnesses, but also to give an error-free judgment of whether a defendant or witness is coming clean.
To resolve court backlogs in Estonia, a robot judge powered by AI analyzes legal documents and other relevant information, and comes to a decision.
What’s the chance, let’s say, especially of a loudmouth run-of-the-mill lawyer/teacher outpacing the ability of this technological innovation that can read, comprehend, analyze, and decide in context millions of legal texts in seconds?
Although in the case of Estonia, the actual decision is left for a human judge to dispense.
As far as education is concerned, I don’t have much to relate of the general direction and quality of our Third World educational system, except what one current graduate told me about a certain university in Dumaguete: “Ngilngig paminawon sa gawas, ug expensive, apan overall, pretentious ug bati ang educational quality.” (Its reputation abroad is excellent/awesome, and it’s an expensive school, but overall, the quality of education is pretentious and substandard.)
Then added, “My parents could have a cheaper and better education for me if I attended the public University nearby. After all, I regard myself as self-motivated and creative. Unsaon man kay mga alumni man gud sila.” (But what can I do since they are alumni?)
Computer is now more than a tool or a compute; it is becoming a sophisticated machine learning agent interacting in the world; and improving itself towards super intelligence.
This is now possible as computers are slowly moving away from classical computers (i.e., smartphones, laptops, or supercomputers) into quantum computers. (Classical computing uses a bit of 0 (off) or a 1 (on), while quantum computing uses a qubit that can be on and off at the same time, a condition known as superposition and entanglement).
For example in 2019, Google ran a calculation on a quantum computer in just a few minutes that would take a classical computer 10,000 years to complete.
A little over a year later, China performed a calculation in 200 seconds that would take an ordinary computer 2.5 billion years — or 100 trillion times faster.
When Deep Mind Technologies’ (a Google subsidiary) Alpha Go beat the world’s best ‘Go’ player, I was not surprised. There is a precedent.
For example, a tech giant like IBM has an engineering history of beating the world’s best players i.e., IBM’s Deep Blue defeated the world’s best chess player, and IBM’s Watson defeated the world’s best Jeopardy quiz show player.
But when Google created Alpha Go Zero, I was more alarmed than thrilled.
While Alpha Go imitates human strategies, Alpha Go Zero, sans data from human games, played and learned against itself, and created strategies alien to humans. In other words, it reached a threshold of deep learning towards super intelligence.
In the coming days, our emerging technological experience will have an increasing impact on our everyday life.
As a sampler of the now in some advanced countries, you may watch Beyond Atlas and Ameca on YouTube.
Inevitably, we have to address these fundamental questions: How do we converge the decision-making of highly-advanced softwares to a state that is aligned with human interest? How do we code this shared value of alignment?
For me, the resolution of these questions will either determine the continuation of the human era, or the end of it.
______________________________________
Author’s email: [email protected]u
