-
by Josh Ong
, Op-Ed Contributor,
January 22, 2018
Artificial intelligence is a polarizing topic. Some, like Elon Musk, think it has the capacity to ruin the world;
others, like Mark Zuckerberg, are optimistic about the benefits it will bring. It often feels like there’s no middle ground between the two: you either believe in AI, or you don’t.
However, regardless of your stance on the topic, there is no denying that we are currently in an AI boom, fueled by the extraordinary strides in machine intelligence made in the past several years. That being said, we are also “a long way from artificial general intelligence”, says Erik Brynjolfsson, a professor at MIT’s Sloan School of Management
and the author of a recent report attempting to measure how far along we are in our attempts to create human-level machine intelligence. Brynjolfsson points out that while there have been considerable
advances in certain fields of AI, especially with regard to image and voice recognition, computers that are trained to perform those tasks are generally unable to do much else, and are also unable to
“adapt if the nature of the task changes slightly or if they see something completely unfamiliar.” To say that AI is just hype is to ignore the not-insignificant breakthroughs that have been made in
the field, breakthroughs that are currently living in your smartphone and computer. It also ignores the fact that not every AI has to have human-level intelligence in order to carry out its
tasks.
We also have to
separate the technology itself from the claims that companies make in order to sell their products. Like Big Data before it, AI has become such a hot trend that startups are tempted to overstate their
AI capabilities to attract interest and investment. As Tom Rikert, a partner at Next World Capital,
notes, we’re currently headed toward “a necessary recalibration of the space,” which “will separate the winning AI-driven companies from all the noise” and
promote “substance over flash.”
The other thing to keep in mind is that big tech companies - Google, Facebook, Amazon, and others - have been working on AI for years, and have also acquired a good many AI startups
to help beef up their own AI capabilities. First and foremost amongst these acquisitions is DeepMind, purchased by Google in 2014, which has been making headlines recently for its development of a
deep learning program capable of
teaching itself to play
chess and other strategically complicated games in a matter of hours. The incredible feats of DeepMind’s AlphaZero and IBM’s Watson show just how sophisticated AI can get - and, in
select cases, how sophisticated it already is.
What will mark the turning point from hypothetical to practical? It’s difficult to tell, because we often seem to forget that artificial intelligence is a form
of science, and one that requires significant amounts of experimentation and trial and error before breakthroughs can be achieved. Right now, using AI can be a bit of a hassle, because of the amount
of oversight required; even an AI that has a 90 percent success rate requires a human to supervise its work and correct the 10 percent that’s incorrect.
The answer to my question of whether or not AI is just hype
is complicated. The current boom is in some ways the result of companies inflating the abilities of their products, but there are also many companies out there doing extraordinary work. Ultimately, AI
has a ways to go, but the advances already out there: advanced driver-assistance systems, facial recognition, voice assistants - are proof enough of the incredible potential that AI has to transform
our lives and the way we work.
If people can learn, why cannot machines?