“Should we ____ away all the jobs, including the ____ ones? Should we develop ____ minds that might ____ outnumber, ____...and replace us? Should we risk loss of control of our civilisation?” These questions were asked last month in an open letter from the Future of Life Institute, an ngo. It called for a six-month “pause” in the creation of the most ____ forms of artificial intelligence (ai), and was signed by tech ____ including Elon Musk. It is the most ____ example yet of how ____ progress in ai has ____ anxiety about the potential ____ of the technology. In particular, new “large language models” (llms)—the sort that powers Chatgpt, a chatbot made by Openai, a startup—have surprised even their creators with their ____ talents as they have been ____. Such “____” abilities include everything from solving logic ____ and writing computer code to identifying films from ____ summaries written in emoji. These models stand to transform humans’ relationship with computers, knowledge and even with themselves. ____ of ai argue for its potential to solve big problems by developing new drugs, designing new materials to help fight climate change, or ____ the complexities of fusion power. To others, the fact that ais’ ____ are already ____ their creators’ understanding risks bringing to life the science-fiction disaster scenario of the machine that ____ its inventor, often with ____ consequences. This ____ mixture of excitement and fear makes it hard to ____ the opportunities and risks. But lessons can be learned from other industries, and from past technological ____. So what has changed to make ai so much more capable? How ____ should you be? And what should governments do? In a special Science section, we ____ the workings of llms and their future direction. The first wave of modern ai systems, which emerged a decade ago, ____ on carefully labelled training data. Once ____ to a ____ number of labelled examples, they could learn to do things like ____ images or ____ speech. Today’s systems do not require pre-labelling, and as a result can be trained using much larger data ____ taken from online ____. llms can, in effect, be trained on the ____ internet—which explains their capabilities, good and bad. Those capabilities became ____ to a wider public when Chatgpt was released in November. A million people had used it ____ a week; 100m within two months. It was soon being used to ____ school essays and wedding speeches. Chatgpt’s popularity, and Microsoft’s move to incorporate it into Bing, its search engine, ____ rival firms to ____ chatbots too. Some of these produced strange results. Bing Chat suggested to a journalist that he should leave his wife. Chatgpt has been accused of ____ by a law professor. llms produce answers that have the ____ of truth, but often contain factual ____ or ____ fabrications. Even so, Microsoft, Google and other tech firms have begun to ____ llms into their products, to help users create documents and ____ other tasks. The recent ____ in both the power and ____ of ai systems, and growing ____ of their abilities and defects, have raised fears that the technology is now advancing so quickly that it cannot be safely controlled. Hence the call for a pause, and growing ____ that ai could ____ not just jobs, factual ____ and reputations, but the existence of humanity itself. Extinction? ____? The fear that machines will steal jobs is centuries old. But so far new technology has created new jobs to replace the ones it has destroyed. Machines ____ be able to perform some tasks, not others, increasing demand for people who can do the jobs machines cannot. Could this time be different? A sudden ____ in job markets cannot be ____, even if so far there is no sign of one. Previous technology has tended to replace unskilled tasks, but llms can perform some ____ tasks, such as summarising documents and writing code. The degree of ____ risk posed by ai has been ____ debated. Experts are divided. In a survey of ai researchers ____ in 2022, 48% thought there was at least a 10% chance that ai’s ____ would be “extremely bad (eg, human extinction)”. But 25% said the risk was 0%; the ____ researcher put the risk at 5%. The ____ is that an advanced ai causes harm on a massive scale, by making poisons or viruses, or ____ humans to ____ terrorist acts. It need not have evil ____: researchers worry that future ais may have goals that do not ____ with those of their human creators. Such scenarios should not be ____. But all involve a huge amount of ____, and a ____ from today’s technology. And many imagine that future ais will have ____ access to energy, money and computing power, which are real ____ today, and could be denied to a ____ ai in future. Moreover, experts tend to ____ the risks in their area, compared with other ____. (And Mr Musk, who is launching his own ai startup, has an interest in his rivals downing tools.) ____ heavy regulation, or indeed a pause, today seems an ____. A pause would also be ____. Regulation is needed, but for more ____ reasons than saving humanity. Existing ai systems raise real concerns about ____, privacy and intellectual-property rights. As the technology advances, other problems could become apparent. The key is to balance the promise of ai with an ____ of the risks, and to be ready to ____. So far governments are taking three different ____. At one end of the ____ is Britain, which has proposed a “light-touch” approach with no new rules or regulatory bodies, but ____ existing regulations to ai systems. The aim is to ____ investment and turn Britain into an “ai superpower”. America has taken a similar approach, though the Biden administration is now ____ public views on what a ____ might look like. The eu is taking a ____ line. Its proposed law categorises different uses of ai by the degree of risk, and requires increasingly ____ monitoring and ____ as the degree of risk rises from, say, music-recommendation to self-driving cars. Some uses of ai are ____ altogether, such as ____ advertising and remote biometrics. Firms that break the rules will be ____. For some critics, these regulations are too ____. But others say an even ____ approach is needed. Governments should ____ ai like medicines, with a dedicated regulator, strict testing and pre-approval before public release. China is doing some of this, requiring firms to register ai products and ____ a security review before release. But safety may be less of a motive than politics: a key requirement is that ais’ ____ reflects the “core value of socialism”. What to do? The light-touch approach is ____ to be enough. If ai is as important a technology as cars, planes and medicines—and there is good ____ to believe that it is—then, like them, it will need new rules. Accordingly, the eu’s model is ____ to the mark, though its classification system is ____ and a principles-based approach would be more ____. Compelling ____ about how systems are trained, how they operate and how they are monitored, and requiring inspections, would be ____ to similar rules in other industries. This could allow for ____ regulation over time, if needed. A dedicated regulator may then seem ____; so too may intergovernmental ____, similar to those that ____ nuclear weapons, should ____ evidence emerge of existential risk. To monitor that risk, governments could form a body modelled on cern, a particle-physics laboratory, that could also study ai safety and ethics—areas where companies ____ incentives to invest as much as society might wish. This powerful technology ____ new risks, but also offers ____ opportunities. Balancing the two means ____ carefully. A measured approach today can provide the ____ on which further rules can be added in future. But the time to start building those foundations is now. ■

How to worry wisely about artificial intelligence

I-print
Naka-embed

Leaderboard

Estilo ng visual

Mga pagpipilian

Magpalit ng template

Ibalik ng awtomatikong pag-save: ?