Christians and artificial intelligence

June 27th, 2017

Playing games with AI

How relevant is a 3,000-year-old board game to humanity’s future? More relevant than you might think, especially when a computer learns how to win.

In May, AlphaGo, an artificial intelligence (AI) developed by Google’s DeepMind Technologies, defeated world Go champion Ke Jie. Go is renowned for its complexity — it has over 300 times the number of possible plays as chess — and requires successful players to utilize artful intuition. Whereas AlphaGo first defeated a human champion in 2016, it won this year’s victory while using ten times less computational power. Ke told The Guardian that AlphaGo has “become too strong for humans,” saying, “I feel like his game is more and more like the ‘Go god.’ ”

Other DeepMind AIs were recently tasked to play a game that involved gathering “apples” (green pixels on a screen). As the apples grew scarce, the AIs learned to use the “highly aggressive” tactic of shooting simulated lasers at each other. In another game, they adapted and learned to use cooperative behavior, working together as “wolves” to capture “prey.”

Game-playing AIs offer an intriguing glimpse into humanity’s future relationship to machines. Some observers expect that this future will be filled with human-AI partnerships that enhance our lives, societies and the planet as a whole. Others warn that AIs could turn on us, as they turned on each other when gathering apples, and defeat us as decisively as the AlphaGo defeated Ke Jie.

What is artificial intelligence?

In 1955, John McCarthy, an assistant mathematics professor at Dartmouth College, proposed a conference to research “the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.” The conference would study “how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves.” McCarthy referred to the aim of this research as “artificial intelligence.” 

The Dartmouth Conference agenda charted the course for the AI research that continues around the world today. Here are a few examples from recent headlines:

• In China, the AI-MATHS machine finished the math section of that nation’s university entrance exam in 22 minutes. By comparison, human students have up to two hours. However, the machine didn’t outperform the human students scoring 100 of 150 points compared to last year’s human average of 109.

• The computers of Google’s Project Magenta sketch original artwork that corrects the intentionally flawed examples of human artists. “If you draw a pig with eight legs,” explains lead researcher Douglas Eck, “the [computer’s] reconstruction model actually gives it four legs because it knows that pigs have four legs.”

Researchers aren’t the only ones interacting with AI. You might be, too — by some definitions, at least. Maybe you use iPhone’s Siri program or Amazon’s Alexa to order merchandise, play music, make dinner reservations and more. Maybe you choose TV shows or movies to watch based on recommendations made by the algorithms of Netflix or Hulu. You might even use a thermostat that learns your daily routines and temperature preferences to self-program energy-efficient heating and cooling schedules.

Writing for Wired, computing expert Assaf Baciu points out that currently “no existing AI technologies can master even the simplest challenges without human-provided context.” Even so, based on these early steps, many experts believe an AI that does more than “simulate” intelligence is within reach.

The emergence of true AI would affirmatively answer the question pioneering computer scientist Alan Turing asked in 1950: “Can machines think?” A concept once regarded as science fiction could soon be fact, and machine minds could be as self-directed and self-aware as the mortal minds that made them.

“Summoning the Demon”

Science fiction has a long history of conjuring threatening scenarios based on AI. Who can forget the psychotic, murderous HAL from 2001: A Space Odyssey or the human-enslaving machines of The Matrix? However, near-term threats are a more immediate cause for concern. Speaking about Tesla’s own efforts to develop AI for self-driving cars, Elon Musk said that self-driving cars will “happen much faster than people realize,” and society must be ready to help the large numbers of people who work as drivers because they will be left unemployed.

Some experts fear that AI may pose more serious dangers. Computer science professor Vernor Vinge has popularized the idea of the Singularity, the moment when we “create intelligences greater than our own . . . [and] the world will pass far beyond our understanding.” Stephen Hawking contends that true AI could “take off on its own, and re-design itself at an ever increasing rate. . . . The development of full artificial intelligence could spell the end of the human race.”

Earlier this year, many leading AI researchers and investors pledged support for guidelines to ensure that AI will offer “amazing opportunities to help and empower people in the decades and centuries ahead.” Google researchers propose a “big red button” humans could use to stop AI from doing harm. But will such precautions be enough? “We might include fail-safes and off-buttons,” argues neuroscientist Dr. Yohan John, but “algorithms might find ways to circumvent them. . . . And if the [malfunctioning] device is a military drone or a future soldier-bot, then we may be in serious trouble!”

Musk told an MIT audience in 2014, “With artificial intelligence, we are summoning the demon.” He explained that it’s like in the stories you hear of the guy who’s absolutely certain he can “control the demon” using whatever means are available to him, but in the end he can’t.

Problem-solving partners

Many researchers and proponents are confident that AI isn’t the unwritten horror story of humanity’s future. Far from causing problems, they contend that thinking machines might solve them.

For example, another AI designed by Google’s DeepMind team learned how to cut the electricity used to cool Google’s data centers by 40 percent. Given that data centers demand three percent of the world’s annual energy, this system has huge economic and environmental benefits.

Health care is another real-world arena where AI could contribute dramatically. Researchers at Houston Methodist Hospital have developed software that interprets mammogram results 30 times faster than human physicians, and with 99 percent accuracy. According to the Houston Chronicle, “Human review . . . results in unnecessary biopsies nearly 20 percent of the time.” This use of AI saves not only money and time, but lives.

Microsoft CEO Satya Nadella offered this optimistic vision for the shape of things to come: “Ultimately, humans and machines will work together — not against one another. Computers may win at games, but imagine what’s possible when human and machine work together to solve society’s greatest challenges like beating disease, ignorance, and poverty.”

Thinking theologically about AI

Scripture offers multiple warnings against placing ultimate trust in anything but God. Relatedly, it documents our tendency to make idols of our achievements and ourselves. However, faith grounded in Jesus’ resurrection gives us a fundamental hope for the future and the belief that God will “bring all things together in Christ” (Ephesians 1:10).

Wouldn’t the God who is sovereign over nations and history be sovereign over AI as well? Couldn’t God use it to fulfill God’s plan? And if that’s the case, couldn’t we partner with AI to love our neighbors better and care for God’s world in new, faithful ways?

The Reverend Dr. Christopher Benek, a Presbyterian pastor who studies emerging technologies, thinks we can. “When we pray ‘on earth as it is in heaven,’” he told me in a 2015 interview, “I think that actually means something. . . . We have the opportunity, like never before, to make the world better. . . . Technology is a tool, and we have a responsibility as Christians to try to use it for good. . . . We need to learn what technology is coming about and advocate in ways that are just and [that align with] the purposes of Jesus.” 

Be sure to check out FaithLink, a weekly downloadable discussion guide for classes and small groups.

comments powered by Disqus