Discover more from Last Week in AI
Skynet Today Last Week in AI News #49
Last Week in AI News #49
Subscribe for future emails here!
Mini Briefs
The Secretive Company That Might End Privacy as We Know It
Clearview AI has been in the news lately for its controversial facial recognition service, which it sells to law enforcement agencies mostly in the U.S. Its product is popular among the police because 1) Clearview has amassed a database of more than 3 billion public images of people’s faces from sites like Facebook, Twitter, and Venmo (this scraping violates those companies’ terms of use) and 2) it uses the latest advances from Big Ol' Neural Nets to perform the recognition, so the input image does not have to be taken perfectly head-on at eye level.
While Clearview has helped to identify many suspects and catch criminals, its effectiveness and scale is a growing concern. For one, the company keeps all the photos that law enforcement agencies upload, and for another there is evidence to suggest that it is keeping track of these queries. The potential for abuse is also alarming: “Imagine a rogue law enforcement officer who wants to stalk potential romantic partners, or a foreign government using this to dig up secrets about people to blackmail them or throw them in jail.” Clearview may also open the application to the public, and even if they don’t, another company might try. With such a tool:
Searching someone by face could become as easy as Googling a name. Strangers would be able to listen in on sensitive conversations, take photos of the participants and know personal secrets. Someone walking down the street would be immediately identifiable — and his or her home address would be only a few clicks away. It would herald the end of public anonymity.
Artificial Intelligence Needs Private Markets for Regulation - Here’s Why
To help government regulations catch up with the fast, evolving field of AI and to also avoid stifling innovation with too much control, Jack Clark and Gilian Hadfield from OpenAI propose the idea of “regulatory markets.” Essentially, to regulate AI applications in the private sector, government would issue licenses to third-party AI regulatory companies that actually perform the regulation, from checking in on AI companies to collecting fines.
One example of such a regulatory market is credit agencies, but to maintain the integrity and effectiveness of the regulation service providers, the regulatory market must be kept competitive. The government must be ready to withdraw licenses and funding to disincentivize complacent behavior.
A regulatory market approach enables the dynamism needed for AI to flourish in a way consistent with safety and public trust. The speed of private innovation should not make governments reluctant to act. Instead, policy needs to get innovative too.
Advances & Business
The Economics of AI Today - Every day we hear claims that Artificial Intelligence (AI) systems are about to transform the economy, creating mass unemployment and vast monopolies. But what do professional economists think about this?
Inside big tech’s quest for human-level A.I. - Giants like Alphabet and Microsoft are investing heavily to develop technology that could radically reshape the business world. When it arrives is anyone’s guess.
IBM’s debating AI just got a lot closer to being a useful tool - Computers have guided us to the moon and back but can’t help us with us with the biggest decisions we face today.
How Advances in AI and Robotic Surgery Will Affect Nurses and APPs - The use of artificial intelligence (AI) and robotics in surgery has been growing for several decades. During the last 14 years in the United States, patients have undergone 1.75 million robotic surgical procedures.
Robot maker Boston Dynamics replaces CEO to prepare for “new stage of growth” - Boston Dynamics, the creator of the internet’s favorite robot dog, is announcing a new CEO and developer conference as it begins selling its robots more widely.
DeepMind researchers introduce hybrid solution to robot control problems - Fundamental problems in robotics involve both discrete variables, like the choice of control modes or gear switching), and continuous variables, like velocity setpoints and control gains.
GitHub now uses AI to recommend open issues in project repositories - Large open source projects on Github have intimidatingly long lists of problems that require addressing.
Amazon explores a way to preserve privacy in natural language processing - Can privacy and security be preserved in the course of a large-scale data analysis? As it turns out, yes. A team of Amazon researchers in a recently published study propose a way to anonymize customer-supplied textual data.
bjork and microsoft use AI to create music that changes with the weather - bjork has collaborated with microsoft to create AI-generated music based on on the changing weather patterns and position of the sun.
Facebook speeds up AI training by culling the weak - Training an artificial intelligence agent to do something like navigate a complex 3D world is computationally expensive and time-consuming.
Waymo commits to more safety-drivers for its ‘driverless’ cars - Alphabet Inc.’s Waymo took a big step forward last fall when it began ferrying riders around the Phoenix, Arizona, area in robotaxis without human safety drivers. Humans have been behind the wheel for almost all of the 20 million miles of testing the company says it’s completed on public roads.
GM’s Cruise has unveiled a self-driving car with no steering wheel - Cruise, the autonomous-car division of General Motors, has unveiled a futuristic electric vehicle called the Origin that completely does away with a human driver. It isn’t intended for private use.
AI Helps Spot Dental Fraud - System from startup Pearl Inc. flags potentially fraudulent claims, administrative errors.
Artificial Intelligence for Breast Cancer Detection - Researchers at the NYU Center for Data Science and the NYU School of Medicine have created an AI diagnostic tool that’s helping doctors detect breast cancer with more accuracy.
Concerns & Hype
Why using AI to screen job applicants is almost always a bunch of crap - Millions of potential employees are subjected to artificial intelligence screenings during the hiring process every month. While some systems make it easier to weed out candidates who lack necessary educational or work qualifications, many AI hiring solutions are nothing more than snake oil.
There’s a new obstacle to landing a job after college: Getting approved by AI - College career centers used to prepare students for job interviews by helping them learn how to dress appropriately or write a standout cover letter. These days, they’re also trying to brace students for a stark new reality: They may be vetted for jobs in part by artificial intelligence.
Amazon asks court to halt Microsoft’s work on Pentagon ‘war cloud’ - Amazon on Wednesday asked a U.S. federal court to stop Microsoft from working with the Pentagon to implement a $10 billion cloud-computing contract, arguing that the project should stall until the courts work out whether Microsoft deserved to receive the lucrative deal.
Microsoft Looms Over the Privacy Debate in Its Home State - The software company helped torpedo a facial recognition bill last year, though a state senator—who’s also a Microsoft program manager—has a new bill in the works.
Google owner calls for ‘proportionate approach’ to AI regulation - Sundar Pichai, the CEO of Google-owner Alphabet, urged regulators on Monday to take a “proportionate approach” when drafting rules for artificial intelligence (AI), days before the European Commission is due to publish proposals on the issue.
The EU might ban facial recognition in public for five years - The news: The European Commission is considering a ban of facial recognition in public places for up to five years, with exceptions for research and security projects, according to a white paper draft obtained by Politico. (The paper does not specify what these projects might be.
When bias begets bias: A source of negative feedback loops in AI systems - Is bias in AI self-reinforcing? Decision-making systems that impact criminal justice, financial institutions, human resources, and many other areas often have bias. This is especially true of algorithmic systems that learn from historical data, which tends to reflect existing societal biases.
AI Can Do Great Things - If It Doesn’t Burn the Planet - The computing power required for AI landmarks, such as recognizing images and defeating humans at Go, increased 300,000-fold from 2012 to 2018.
Researchers use AI to deblur human faces in photos - Scientists at the Inception Institute of Artificial Intelligence in the United Arab Emirates, the Beijing Institute of Technology, and Stony Brook University developed an AI system that removes blur from images in post-production. They note in a paper that it’s human-aware, meaning it’s able to deblur human faces, and that it performs “favorably” against state-of-the-art motion deblurring methods.
How a Kaggle Grandmaster cheated in $25,000 AI contest with hidden code - and was fired from dream SV job - A Google-backed competition to develop machine-learning software to help abandoned animals find loving homes turned ugly when it was revealed the winning team cheated.
Met police to begin using live facial recognition cameras - The Metropolitan police will start using live facial recognition, Britain’s biggest force has announced.
Robot tanks: On patrol but not allowed to shoot - In 1985 the US pulled the plug on a computer-controlled anti-aircraft tank after a series of debacles in which its electronic brain locked guns onto a stand packed with top generals reviewing the device.
The battle for ethical AI at the world’s biggest machine-learning conference - Bias and the prospect of societal harm increasingly plague artificial-intelligence research — but it’s not clear who should be on the lookout for these problems.
Key Volkswagen Exec Admits Full Self-Driving Cars ‘May Never Happen’ - Full autonomy remains an impossibly difficult prize. After years of sunny optimism and big promises, automakers are beginning to realize just how difficult it is to make a market-ready, full self-driving car.
Analysis & Policy
Patterns of Life: AI and “Actionable Data” in Warfare - The past two decades have brought two interrelated and disturbing developments in the technopolitics of US militarism. The first is the fallacious claim for precision and accuracy in the United States’ counterterrorism program, particularly for targeted assassinations.
RSNA 2019 AI round-up - Hardian Health - Scientific support for structured reporting was also a theme in the academic sessions.
We’re Banning Facial Recognition. We’re Missing the Point. - The whole point of modern surveillance is to treat people differently, and facial recognition technologies are only a small part of that.
The AI delusion: why humans trump machines - As well as playing a key role in cracking the Enigma code at Bletchley Park during the Second World War, and conceiving of the modern computer, the British mathematician Alan Turing owes his public reputation to the test he devised in 1950.
Explainers
The Evolution of AlphaGo to MuZero - DeepMind recently released their MuZero algorithm, headlined by superhuman ability in 57 different Atari games.
Emil’s Story as a Self-Taught AI Researcher - Emil is living, breathing proof that it’s possible to pursue serious AI research as a self-taught creator.
Is MT really lexically less diverse than human translation? - What got me interested in lexical diversity of MT outputs? This blog post describes my journey into the rabbit hole that is lexical diversity of machine translation outputs.
Table Detection, Information Extraction and Structuring using Big Ol' Neural Nets - The amount of data being collected is drastically increasing day-by-day with lots of applications, tools, and online platforms booming in the present technological era. To handle and access this humongous data productively, it’s necessary to develop valuable information extraction tools.
The Next Generation of Machine Learning Tools - ML research is starting to hit the limits of the tools we currently have at our disposal, but some new and exciting candidates are right around the corner, such as JAX and S4TF.
Black-Boxed Politics: - Opacity is a Choice in AI Systems