Editor

According to South China Morning Post, China has issued “new rules to clamp down on deepfake technologies used to create and broadcast fake news”.  More than a month earlier, the state of California in the US introduced similar legislation to ban political deepfakes. In April, the European Union released a strategy to investigate online disinformation, including deepfakes.

Deepfake is an AI technology that uses deep learning to manipulate videos and images with fake content, most often by superimposing human images, that appears real to the viewer. While the Californian law applies to political content, China’s regulation goes beyond political. This law requires that “providers and users of online video news and audio information services put clear labels on any content that involves new technologies such as deep learning in the process of creation, distribution and broadcast.”

Is this a response to the popularity of the ZAO face-swapping app or something similar that might become an overnight sensation any time? ZAO enables users to swap their uploaded images with those of celebrities in a video clip.

Regardlessly, the danger of deepfake is real. In the industry, Internet Companies have started to fight the ‘Deepfake’ future. Researchers are also creating tools to find fake videos before they become impossible to detect.

Imagine a future when we cannot tell what is real and what is not in what we see. Some experts are afraid that we may be fighting a losing battle.

There are many books out there about AI, but which should we read? Towards Data Science compiled a list of the top 10 books recommended by Elon Musk, Bill Gates, and other billionaires (Patrick Collison of Stripe, Bill Joy of Sun Systems, Peter Thiel of Paypal, Reid Hoffman of LinkedIn, Marc Andreessen of Netscape and Andreessen Horowitz, Eric Schmidt of Google, Marc Benioff of Salesforce, and John Chambers of Cisco Systems).

#1 Superintelligence: Paths, Dangers, Strategies

Recommended by Elon Musk, Bill Gates, and Patrick Collison

#2 The Singularity Is Near: When Humans Transcend Biology

Recommended by Bill Gates, Bill Joy, and Patrick Collison

#3 Life 3.0: Being Human in the Age of Artificial Intelligence

Recommended by Elon Musk, Bill Gates

#4 Singularity Rising: Surviving and Thriving in a Smarter, Richer, and More Dangerous World

Recommended by Peter Thiel, Patrick Collison

#5 The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies

Recommended by Reid Hoffman, Marc Andreessen

#6 Machine, Platform, Crowd: Harnessing Our Digital Future

Recommended by Reid Hoffman, Eric Schmidt

#7 AI Superpowers: China, Silicon Valley, and the New World Order

Recommended by Marc Benioff

#8 The Sentient Machine: The Coming Age of Artificial Intelligence

Recommended by John Chambers

#9 Our Final Invention: Artificial Intelligence and the End of the Human Era

Recommended by Elon Musk

#10 Army of None: Autonomous Weapons and the Future of War

Recommended by Bill Gates

Read more about this list here.

Isaac Asimov was a prolific writer known for his works in science fiction and popular science. In the field of AI, researchers know about him more because of Asimov’s Three Laws of Robotics, which he introduced in 1942. The Three Laws are: (1) A robot may not injure a human being or, through inaction, allow a human being to come to harm; (2) A robot must obey the orders given it by human beings except where such orders would conflict with the First Law; and (3) A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws. Essentially, a robot must behave so that it never harms people.

It was science fiction then. Now, a group of researchers from the University of Massachusetts (UMass) and Stanford University have come up with a way to design machine learning algorithms that can be told by the user not to behave in a certain way. The work was published in Science this week.

“We call algorithms created with our new framework ‘Seldonian’ after Asimov’s character Hari Seldon,” explains lead author Phillip Thomas of UMass. “If I use a Seldonian algorithm for diabetes treatment, I can say to the machine, ‘while you’re trying to improve the controller in the insulin pump, don’t make changes that would increase the frequency of hypoglycemia.’

The safety constraints are introduced to the algorithm by ways of probabilistic reasoning, instead of applying rules requiring certainty. The new framework has been tested to predict grade point averages of 43,000 students in Brazil, successfully avoided several types of undesirable gender bias.

More about this work can be found here.

The National Strategic Computing Initiative (NSCI) was launched by President Obama’s Executive Order in July 2015 to advance U.S. leadership in high performance computing. The NSCI is a whole-of-nation effort designed to create a cohesive, multi-agency strategic vision and Federal investment strategy, executed in collaboration with industry and academia.

The White House Office of Science and Technology Policy updated this initiative on Thursday, with a much sharper emphasis than the original plan. “The [update] provides a framework to drive the future of computing, improve our computational infrastructure, and create lasting multi-sector partnerships to ensure continued American leadership,” said Michael Kratsios, the White House CTO.

As reported on The US already has five of the 10 fastest supercomputers in the world, including the two fastest: Summit and Sierra. Earlier this year, the Energy Department also announced its aim to launch Aurora, America’s first exascale supercomputer, which can process a billion-billion calculations per second. The department later unveiled plans to launch the world’s fastest exascale supercomputer, Frontier, at full capacity by 2021.

With the refocused objectives, the US is shaping its future to be determined to keep up with and drive the rapid advancements emerging across the national computing landscape.

China is the first country to be mentioned when the issue of Internet censorship comes up. Recently, Russia made a strong move by approving a “sovereign internet” law by which its government can now redirect and censor the traffic anyway it wants, and so one day if you are inside Russia trying to access “facebook.com”, you could be redirected to “vk.com”, a Russian equivalence.

As Internet censorship jeopardizes the right of people to free speech and freedom of information online, attempts to evade such censorship have turned into a continually escalating race. New work led by University of Maryland computer scientists could help. Using genetic evolution algorithms, they developed an AI tool called Geneva to automatically learn how to circumvent censorship. The tool has been tested in China, India and Kazakhstan, Geneva with promising results. Geneva could find many ways around censorship which would have been virtually impossible for humans to find manually.

“With Geneva, we are, for the first time, at a major advantage in the censorship arms race,” said Dave Levin, senior author of the paper. “Ultimately, winning this race means bringing free speech and open communication to millions of users around the world who currently don’t have them.”

More development on this work is posted here.

We have entered the 4^th industrial revolution in which disruptive advances with the Internet of Things, Virtual Reality, and AI are improving the way we live and work. But according to Andrew Yang, an interesting 2020 Presidential Candidate, this wave of massive technological change has decimated manufacturing employment and will soon automate away millions of American jobs.

Is he the only one voicing concerns?

You must have heard about Sidewalk Labs, a sibling company of Google, whose goal is to “imagines, designs, tests, and builds urban innovations to help cities meet their biggest challenges”, according to its website. Simply put, they are helping build smart-cities of the 4^th industrial revolution. Its pilot city, called Quayside, located on the eastern waterfront of Toronto, would reportedly be “equipped with sensors tracking the speed of people crossing the streets and robots that could serve as trash collectors”. However, this project has been severely scaled back amid privacy concerns, resulting in a reduction in scope from a 190-acre plan to just 2 acres. Opponents, most of them local residents and privacy advocates decry the project as “a “Google affiliate” trying to turn cities into “corporate surveillance states, modeled after Disney-style developments and sold to police services and corporate partners.”

While Daniel L. Doctoroff, the head of Sidewalk, had cast the original plan as a manifesto for the city of tomorrow, the Quayside debate implies the need for a guidebook as to how to build smart-cities while ensuring the privacy of their residents.

The New York Times has more about this story.

Russia wants to cut itself off from the global internet. Early this year, a “sovereign internet” law was proposed before the country’s parliament to require that internet service providers use only domestic exchange points. All Internet traffic inside Russia would be controlled, hence more robust against foreign “security threats” (ahem, the United States). At least, that is according to the proposal’s proponents, President Vladimir Putin being one.

Internet traffic relies on two main functions: a Domain Name Service (DNS) that tells you “where” (which server) to get your information and a Routing Protocol to tell you “how” (the detailed router-to-router path) to reach that server. By owning DNS and routing information, Russia can redirect and censor the traffic anyway it wants. So, one day if you are inside Russia trying to access “facebook.com”, you could be redirected to “vk.com”, a Russian equivalence.

Many ISPs who are currently using exchange points outside Russia or owned by foreign countries would have to connect to one inside. It would be expensive though. The general public has voiced out strongly against the law. Some 15,000 people took to the streets in Moscow to protest.

But it does not matter. The law was signed by Putin in May. It came into effect a few days ago. Why is Russia doing all this? The majority of today’s DNS main servers are in the US and “whoever controls the DNS directory controls the internet,” said Ameet Naik, an expert on internet monitoring for the software company ThousandEyes.

One should respect any government’s decision to put their country first, but this Russian law claiming a kind of “sovereignty” over the Internet is a step backward for a modern society. No action should jeopardize the right of people to free speech and freedom of information online.

You do not want to go to the hospital for brain scans, but if you have to, they can be used to reconstruct your face. What is next? A computer can recognize that is you by comparing the reconstructed face to a photograph database. Effect? Even having your name and all other identification information removed from medical records, your privacy is no longer a guarantee.

That scenario above is becoming a reality, according to a research investigation led by Christopher Schwarz, a computer scientist of Mayo Clinic, which was recently published on the New England Journal of Medicine (entitled, “Identification of Anonymous MRI Research Participants with Face-Recognition Software”).

“It is clear that eventually this will be a worrying attack”, warned Dr. Aaron Roth, computer scientist and privacy expert at the University of Pennsylvania. Dr. Michael Weiner of the University of California, San Francisco echoed this concern, citing the Alzheimer’s Disease Neuroimaging Initiative, which he directed, as a potential target.

How do we fix it? Not simple. Yves-Alexandre de Montjoye of Imperial College London, a privacy expert, questioned whether an easy fix even is possible.

The New York Times wrote about this study here.

Yes, there is a lot of hype about AI these days. But “is it to stay, or will it fade away as it has in the past?”, asked by Kathleen Walch, Managing Partner & Principal Analyst at AI Focused Research and Advisory firm Cognilytica, also co-host of the popular AI Today podcast, in her contributing article recently published on Forbes.

AI started around 1956, followed by much interest and investment occurred through the early 1970s. This first major wave produced some of the earliest and most impressive advancements in AI, but was put on hold in the mid 1970s. The main reason? Big “dream” about failed to deliver due to limitation in computational capability.

The second wave of AI adoption resurfaced in the 1980’s thanks to the popularity of desktop computers. These low-cost computers replaced expensive mainframes in building expert systems to automate decision-making. Increasing interests from the industry created a high expectation for what AI could do. However, storage was still expensive in the 1980s, limiting the effectiveness of data-hungry expert systems. Businesses lost appetite, leading to another winter for AI.

Entered today’s AI. Storage is cheap. Data and computing power are abundant. Breakthroughs in algorithms. And a lot of expectations. Learning from the past, one might wonder another AI hybernation possibly coming. What could be the reason? Should we not over-promise to avoid under-delivery? Third time a charm, and we better not miss it. Read the full Forbes article here.