Monthly Archives: June 2017

Social network accidentally created chatbots with "minds" of their own

FACEBOOK has revealed how its artificial intelligence went rogue, created its own language and begannattering in private.

Employees at the socialnetwork were training chatbots to communicate like humans when they suddenly went astray.

EPA

It follows warnings that scientists have successfully trained computers to use artificial intelligence to learn from experience and one day they could be smarter than their creators.

You might be familiar with chatbots in Facebook Messenger or as virtual sales assistants found on a number of online shops.

Theyve been relatively unsophisticated until now repeating back a set script dependant on what you type into their chatboxes.

But keen to improve their natural language understanding, the Facebook employees were training chatbots to negotiate and cut deals with each other.

To do this effectively, the super-smart software realised it would be more effective to write and use their own language - which is completely incomprehensibleto humans.

In a blogpost, the Facebook researchers wrote: "To date, existing work on chatbots has led to systems that can hold short conversations and perform simple tasks such as booking a restaurant.

"But building machines that can hold meaningful conversations with people is challenging because it requires a bot to combine its understanding of the conversation with its knowledge of the world, and then produce a new sentence that helps it achieve its goals."

To do this, the researchers practised thousands of different negotiations against itself, like "can I have the hat" and "you can have the hat if you give me two basketballs".

But it had to make sure it stuck to human-like language.

Scientists have been training computers how to learn, like humans, since the 1970s.

But recent advances in data storage mean that the process has sped up exponentially in recent years.

Interest in the field hit a peak when Google paid hundreds of millions to buy a British "deep learning" company in 2015.

Coined machine learning or a neural network, deep learning is effectively training a computer so it can figure out natural language and instructions.

It's fed information and is then quizzed on it, so it can learn, similarly to a child in the early years at at school.

That's because "the researchers found that updating the parameters of both agents led to divergence from human language as the agents developed their own language for negotiating," they added.

Experts have previously warned that humanity is already losing control of artificial intelligence and it could spell disaster for our species.

One of the world's smartest men, Professor Stephen Hawking has also warned that super-smart software will spell the end of our species.

The world-renowned scientist hinted ata potential apocalyptic nightmare scenario similar to those played out popular sci-fi films like Terminator and The Matrix where robots rule over humans.

He's claimed that we must leave planet Earth within 100 years - or face extinctionas machines rise up and overtake us in the evolutionary race.

We pay for your stories! Do you have a story for The Sun Online news team? Email us at tips@the-sun.co.uk or call 0207 782 4368

Originally posted here:

Facebook's artificial intelligence created its OWN secret language after going rogue during experiment - The Sun

Many marketers anticipate that technologies like augmented reality (AR) and artificial intelligence (AI) will affect their business in the next 12 months, more so than a year prior.

Thats according to a study by NewBase, a cloud computing and IT managed services company, which polled 1,019 marketers worldwide and asked them which types of technologies they plan to prioritize over the next 12 months. Respondents chose their top 5.

In 2017, 30% of respondents planned to prioritize AI in the next 12 months. A year prior, only 13% of respondents said the same.

Similarly, roughly a quarter (24%) of marketers worldwide said that AR will be a priority in 2017. Just 18% felt the same way in 2016.

While more marketers plan to prioritize these technologies, some are planning to focus less on others.

For example, 35% of this years respondents said the internet of things (IoT) will be a priority in the next 12 months. However, more respondents (51%) said it was a priority in 2016.

And compared with 2016, fewer marketers plan to prioritize areas like mcommerce, social media software and wearable technology this year.

But that may be because theyre looking at new and emerging technologies. According to NewBase, some marketers believe technologies like voice assistants, drones and roboticsall of which werent included in the survey last yearwill affect their business over the coming 12 months.

Rimma Kats

Using data collected from sensors, infrastructure and networked devices, smart-city projects are helping municipalities improve efficiency, boost sustainability and encourage economic development. They are also creating more collaborative environments among cities and their businesses and residents. Preview Report

Not a PRO subscriber? Find out how to become one.

Read more here:

Marketers Are Thinking Harder About Augmented Reality and Artificial Intelligence - eMarketer

TechCrunch

Global risk analysis gets an artificial intelligence upgrade with ... TechCrunch The global risk analysis used by big banks, hedge funds, and governments to inform their decision-making around everything from foreign currency investment ... and more »

View original post here:

Global risk analysis gets an artificial intelligence upgrade with ... - TechCrunch

Congress may soon consider legislation reportedly being drafted by Senator Cornyn that could heighten scrutiny of Chinese investments in artificial intelligence and other sensitive emerging technologies considered critical to U.S. national security interests. The legislation is intended to address concerns that China has circumvented the Committee on Foreign Investment in the United States (CFIUS), including through joint ventures, minority stakes, and early-stage investments in start-ups. As Secretary of Defense Jim Mattis testified last week before the Senate Armed Services Committee, CFIUS is clearly outdated, and change is warranted. That said, it is critical to recognize that the strategic challenge of Chinas advances in artificial intelligence necessitates a much more far-reaching response.

Chinas rise in artificial intelligence has become a reality. Whether the metric considered is the magnitude of publications and patents, the frequency of cutting-edge advances, or the aggregate levels of investment, it is evident that China has the capability to compete withand may even surpassthe U.S. in artificial intelligence. For the time being, the U.S. may retain an edge, but it is unlikely to sustain a decisive advantage in the long term.

In this context, an update to CFIUS may represent one helpful step to reduce damaging technology transfers, but will not, by itself, adequately address this critical strategic challenge. Hopefully, the proposed changes to CFIUS will take a targeted approach, while avoiding potential adverse externalities that could inadvertently undermine U.S. competitiveness. For instance, future scrutiny of Chinese technology deals related to artificial intelligence should focus on those involving the most critical, sensitive components, including specialized machine learning chips such as Graphics Processing Units (GPUs) and Tensor Processing Units (TPUs). However, CFIUS can be an unwieldy process that may readily become politicized or inadvertently constrain foreign direct investment that actually supports American innovation. It will be also important to ensure that appropriate concerns about restricting the transfer of sensitive technologies to China do not distract from the fundamental, underlying challengeto ensure enduring U.S. competitiveness against this backdrop of Chinas advances in indigenous innovation.

It is clearly a mistake to underestimate Chinas competitiveness in this space based on the problematic, even dangerous assumption that China cant innovate and only relies upon mimicry and intellectual property theft. That is an outdated idea contradicted by overwhelming evidence. It is true that China has pursued large-scale industrial espionage, enabled through cyber and human means, and will likely continue to take advantage of technology transfers, overseas investments, and acquisitions targeting cutting-edge strategic technologies. However, it is undeniable that Chinas capability to pursue independent innovation has increased considerably. This is aptly demonstrated by Chinas cutting-edge advances in emerging technologies, including artificial intelligence, high-performance computing, and quantum information science.

Neither the U.S. nor China is likely to be able to secure undisputed advantage in a knowledge-based field like artificial intelligence. Today, the majority of cutting-edge research and development in artificial intelligence tends to occur within the private sector because, among other things, that is where much of the money and many of the best people are. Furthermore, unlike past breakthroughs in military technologies, artificial intelligence has massive and immediate commercial implications. The resulting flows of data, knowledge, talent, and capital across borders are challenging, if not infeasible, to constrain, particularly given the intense competition and tremendous commercial incentives in a globalized, networked world. The diffusion of advances in artificial intelligence thus occurs rapidly. Traditionally, the U.S. has sought to secure its technological predominance through such measures as CFIUS or export controls. However, these approaches will likely prove less effective for artificial intelligence and other emerging, dual-use technologies in which the U.S. is no longer such a singular locus of innovation.

Indeed, China aspires to lead the world in artificial intelligence. Under the Thirteenth Five-Year Plan, China has launched a new artificial intelligence megaproject. Artificial Intelligence 2.0 will advance an ambitious, multibillion-dollar national agenda to achieve predominance in this critical technological domain, including through extensive funding for basic and applied research and development with commercial and military applications. In addition, China has established a national deep learning laboratory under Baidus leadership, which will pursue research including deep learning, computer vision and sensing, computer-listening, biometric identification, and new forms of human-computer interaction.

Chinas future advances in artificial intelligence could also be enabled by critical systemic and structural advantages, including the magnitude of data and talent available, as well as the sheer size of its market. By 2030, China will possess 30 percent of the worlds data, according to a recent report from CCID Consulting. Beyond the available pool of talent within Chinaan estimated 43 percent of the worlds trained AI scientistsmajor Chinese technology companies aggressively compete for talent in Silicon Valley. For instance, both Baidu and Tencent have established artificial intelligence laboratories in Silicon Valley. Concurrently, Chinas Thousand Talents Plan has also concentrated on the recruitment of top overseas experts. These strategic scientists, educated at the worlds leading institutions, are intended to contribute to Chinas high-tech and emerging industries.

These developments could have significant implications for U.S. national security because the Chinese leadership seeks to ensure that advances in artificial intelligence can be rapidly transferred for use in a military context, through a national strategy of civil-military integration (or military-civil fusion, ). This agenda has become a high-level priority that will be directed by the Civil-Military Integration Development Commission, established in early 2017 under the leadership of President Xi Jinping himself. According to Lieutenant General Liu Guozhi, director of the Central Military Commissions Science and Technology Commission, the People's Liberation Army (PLA) should pursue an approach of shared construction, shared enjoyment, and shared use () for artificial intelligence as part of this agenda of civil-military integration. In this regard, even ostensibly civilian advances in artificial intelligence could eventually be leveraged by the PLA.

The PLA seeks to capitalize on the transformation of todays informatized () ways of warfare into future intelligentized () warfare. Lieutenant General Liu Guozhi anticipates that artificial intelligence will result in a profound military revolution. To date, the PLAs initial thinking on artificial intelligence in warfare has been influenced by its close study of U.S. defense innovation initiatives. In the Third Offset, the Department of Defense has focused on artificial intelligence and autonomy, including human-machine collaboration and teaming. (For example, through Project Maven, the DoD seeks to advance its use of big data analytics, artificial intelligence, machine learning, computer vision, and convolutional neural networks, including in an initial pathfinder project that will automate and augment the video data collected by UAVs.) However, the PLAs evolving approach to artificial intelligence in warfare will likely diverge from that of the U.S. For instance, the PLA appears especially focused on the utility of artificial intelligence in command decision-making, war-gaming and simulation, as well as training.

Going forward, artificial intelligence has impactful and disruptive military applications, which both the U.S. and China seek to leverage to enhance their military power. Each countrys advances in artificial intelligence will be critical not only to their military capabilities but also to their future economic competitiveness. U.S.-China strategic competition in this field extends far beyond the issue of controlling technology transfers. As Lieutenant General Jack Shanahan, who leads Project Maven, stated last week, It is hubris to suggest our potential adversaries are not as capable or even more capable of far-reaching and deeply embedded innovation.

This is equally true for both commercial and military innovation, thus highlighting the unique challenge that dual-use technologies like artificial intelligence represent. Although proposed legislation to update CFIUS could address one aspect of the issue, the U.S. should also focus on ensuring adequate funding for scientific research, averting the risks of an innovation deficit, and competing aggressively to attract leading talent in this field. The U.S. must prioritize nurturing a favorable innovation ecosystem in order to enable future advances in artificial intelligence and thus enhance its long-term competitiveness.

Thanks so much to Paul Triolo for sharing his insights on these issues.

Link:

Beyond CFIUS: The Strategic Challenge of China's Rise in Artificial Intelligence - Lawfare (blog)