Recent advances in artificial intelligence (AI) have been nothing short of dramatic. AI is transforming nearly every sector of society, from transportation to medicine to defense. So it is worth considering what will happen when it becomes even more advanced than it already is.
The apocalyptic view is that AI-driven machines will outsmart humanity, take over the world, and kill us all. This scenario crops up often in science fiction and is easy enough to dismiss, given that humans remain firmly in control.
However, many AI experts take the apocalyptic perspective seriously, and they are right to do so. The rest of society should as well.
Illustration: Mountain people
To understand what is at stake, consider the distinction between “narrow AI” and “artificial general intelligence” (AGI). Narrow AI can operate only in one or a few domains at a time, so while it might outperform humans in select tasks, it remains under human control.
AGI, by contrast, can reason across a wide range of domains, and thus, could replicate many human intellectual skills, while retaining all of the advantages of computers, such as perfect memory recall. Run on sophisticated computer hardware, AGI could outpace human cognition. It is actually hard to conceive an upper limit for how advanced AGI could become.
As it stands, most AI is narrow. Indeed, even the most advanced current systems have only limited amounts of generality. For example, while Google DeepMind’s AlphaZero system was able to master Go, chess and shogi — making it more general than most other AI systems, which can be applied only to a single specific activity — it has still demonstrated capability only within the limited confines of certain highly structured board games.
Many knowledgeable people dismiss the prospect of advanced AGI. Some, such as Selmer Bringsjord of Rensselaer Polytechnic Institute and Drew McDermott of Yale University, say that it is impossible for AI to outsmart humanity.
Others, such as Margaret Boden of the University of Sussex and Oren Etzioni of the Allen Institute for Artificial Intelligence, say that human-level AI might be possible in the distant future, but that it is far too early to start worrying about it now.
These skeptics are not marginal figures, like the cranks who try to cast doubt on climate-change science. They are distinguished academics in computer science and related fields and their opinions must be taken seriously.
Yet other distinguished academics — including David Chalmers of New York University, Yale University’s Allan Dafoe and Stuart Russell of the University of California, Berkeley, Nick Bostrom of Oxford University, and Roman Yampolskiy of the University of Louisville — do worry that AGI could pose a serious or even existential threat to humanity.
With experts lining up on both sides of the debate, the rest of us should keep an open mind.
Moreover, AGI is the focus of significant research and development (R&D). I recently completed a survey of AGI R&D projects, identifying 45 in 30 countries on six continents.
Many active initiatives are based in major corporations such as Baidu, Facebook, Google, Microsoft and Tencent, and in top universities, such as Carnegie Mellon, Harvard and Stanford, as well as the Chinese Academy of Sciences. It would be unwise to simply assume that none of these projects would succeed.
Another way of thinking about the potential threat of AGI is to compare it with other catastrophic risks. In the 1990s, the US Congress saw fit to have NASA track large asteroids that could collide with Earth, even though the odds of that happening are about one in 5,000 per century.
With AGI, the odds of a catastrophe happening in the upcoming century could be as high as one in 100, or even one in 10, judging by the pace of R&D and the level of expert concern.
The question, then, is what to do about it. For starters, we need to ensure that R&D is conducted responsibly, safely and ethically. This will require increased dialogue between those working in the AI field and policymakers, social scientists and concerned citizens.
Those in the field know the technology and will be the ones to design it according to agreed standards; but they must not decide alone what those standards will be.
Many of the people developing AI applications are not accustomed to thinking about the social implications of their work. For that to change, they must be exposed to outside perspectives.
Policymakers will also have to grapple with AGI’s international dimensions.
The bulk of AGI R&D is carried out in the US, Europe and China, but much of the code is open source, meaning that the work potentially can be done from anywhere. So, establishing standards and ethical ground rules is ultimately a job for the entire international community, although the R&D hubs should take the lead.
Looking ahead, some efforts to address the risks posed by AGI can piggyback on policy initiatives already put in place for narrow AI, such as the new bipartisan AI Caucus launched by US Representative John Delaney. There are many opportunities for synergy between those working on near-term AI risks and those thinking about the long term.
However, regardless of whether narrow AI and AGI are considered together or separately, what matters most is that we take constructive action now to minimize the risk of a catastrophe down the road. This is not a task that we can hope to complete at the last minute.
Seth Baum is the executive director of the Global Catastrophic Risk Institute, a think tank focused on extreme global risks. He is also affiliated with the Blue Marble Space Institute of Science and the University of Cambridge Centre for the Study of Existential Risk.
Copyright: Project Syndicate
When Chinese President Xi Jinping (習近平) sits down with US President Donald Trump in Beijing on Thursday next week, Xi is unlikely to demand a dramatic public betrayal of Taiwan. He does not need to. Beijing’s preferred victory is smaller, quieter and in some ways far more dangerous: a subtle shift in American wording that appears technical, but carries major strategic meaning. The ask is simple: replace the longstanding US formulation that Washington “does not support Taiwan independence” with a harder one — that Washington “opposes” Taiwan independence. One word changes; a deterrence structure built over decades begins to shift.
Taipei is facing a severe rat infestation, and the city government is reportedly considering large-scale use of rodenticides as its primary control measure. However, this move could trigger an ecological disaster, including mass deaths of birds of prey. In the past, black kites, relatives of eagles, took more than three decades to return to the skies above the Taipei Basin. Taiwan’s black kite population was nearly wiped out by the combined effects of habitat destruction, pesticides and rodenticides. By 1992, fewer than 200 black kites remained on the island. Fortunately, thanks to more than 30 years of collective effort to preserve their remaining
After Chinese Nationalist Party (KMT) Chairwoman Cheng Li-wun (鄭麗文) met Chinese President Xi Jinping (習近平) in Beijing, most headlines referred to her as the leader of the opposition in Taiwan. Is she really, though? Being the chairwoman of the KMT does not automatically translate into being the leader of the opposition in the sense that most foreign readers would understand it. “Leader of the opposition” is a very British term. It applies to the Westminster system of parliamentary democracy, and to some extent, to other democracies. If you look at the UK right now, Conservative Party head Kemi Badenoch is
A Pale View of Hills, a movie released last year, follows the story of a Japanese woman from Nagasaki who moved to Britain in the 1950s with her British husband and daughter from a previous marriage. The daughter was born at a time when memories of the US atomic bombing of Nagasaki during World War II and anxiety over the effects of nuclear radiation still haunted the community. It is a reflection on the legacy of the local and national trauma of the bombing that ended the period of Japanese militarism. A central theme of the movie is the need, at
RSS