Neural networks often learn to make predictions that overly rely on spurious correlation existing in the dataset, which causes the model to be biased. While previous work tackles this issue by using explicit labeling on the spuriously correlated attributes or presuming a particular bias type, we instead utilize a cheaper, yet generic form of human knowledge, which can be widely applicable to various types of bias. We first observe that neural networks learn to rely on the spurious correlation only when it is "easier" to learn than the desired knowledge, and such reliance is most prominent during the early phase of training. Based on the observations, we propose a failure-based debiasing scheme by training a pair of neural networks simultaneously. Our main idea is twofold; (a) we intentionally train the first network to be biased by repeatedly amplifying its "prejudice", and (b) we debias the training of the second network by focusing on samples that go against the prejudice of the biased network in (a). Extensive experiments demonstrate that our method significantly improves the training of network against various types of biases in both synthetic and real-world datasets. Surprisingly, our framework even occasionally outperforms the debiasing methods requiring explicit supervision of the spuriously correlated attributes.

The World Economic Forum's Frameworks for Reimagining Regulation at the Age of AI seek to address the need for upgrading our existing regulatory environment to ensure the trustworthy design and deployment of AI. These frameworks provide Governments with innovative approaches and tools for regulating AI that can be scaled.

Award-winning OxFirst, a specialist in the law and economics of IP, has released research that reveals the hidden secrets behind global patent registrations and information on the economic value of patents in the AI sector. While Samsung, IBM and Tencent dominate with the highest number of patents filed, fierce competition between the US and China overshadows other countries, including the EU. Patents are mainly filed in the area of interconnectivity and system architecture, suggesting that top players focus primarily on protecting technologies covering multiple neural nets. Other areas of crucial importance are Machine Learning (ML) and bootstrap methods, alongside procedures used during speech recognition processes; e.g. the further establishment of human-machine dialogue. An analysis of the patent landscape between 2010 and 2020 shows that patents reading on Machine Learning experienced their greatest filing growth in 2017/2018.

The ban comes after civil liberties groups highlighted what they described as faults in facial recognition algorithms after NIST found most facial recognition software was more likely to misidentify people of colour than white people. The Boston ban follows a ban imposed by San Francisco on the use of face recognition technology last year. The ban prevents any city employee using facial recognition or asking a third party to use the technology on its behalf. Boston's police department said it had not used the technology over what it called reliability fears, though it's clear the best systems are reasonably accurate in average working conditions. Critics also opposed the technology on the basis it might discourage citizens' rights to protest.

In 1963, Martin Luther King gave his "I have a dream" speech, words that reflected the thoughts and attitudes of civil rights activists at the time, and lit a torch that lives on in the hearts and minds of those who fight for civil liberties and equality in the western hemisphere. While the world has advanced since Dr. King ushered those words, it's hard to deny that discrimination still rears its ugly head in modern society. We know for a fact that racial discrimination in the workplace is illegal in most of America and Europe. And yet, just in the USA statistics show that things don't seem to have improved regarding hiring practices for black people and Hispanics in the last 25 years. In theory, AI-assisted hiring is built on an underlying model that makes unbiased decisions as long as the data itself isn't biased.

Each Fourth of July for the past five years I've written about AI with the potential to positively impact democratic societies. I return to this question with the hope of shining a light on technology that can strengthen communities, protect privacy and freedoms, or otherwise support the public good. This series is grounded in the principle that artificial intelligence can is capable of not just value extraction, but individual and societal empowerment. While AI solutions often propagate bias, they can also be used to detect that bias. As Dr. Safiya Noble has pointed out, artificial intelligence is one of the critical human rights issues of our lifetimes.

Decision-making on numerous aspects of our daily lives is being outsourced to machine-learning (ML) algorithms and artificial intelligence (AI), motivated by speed and efficiency in the decision process. ML approaches—one of the typologies of algorithms underpinning artificial intelligence—are typically developed as black boxes. The implication is that ML code scripts are rarely scrutinised; interpretability is usually sacrificed in favour of usability and effectiveness. Room for improvement in practices associated with programme development have also been flagged along other dimensions, including inter alia fairness, accuracy, accountability, and transparency. In this contribution, the production of guidelines and dedicated documents around these themes is discussed. The following applications of AI-driven decision-making are outlined: (a) risk assessment in the criminal justice system, and (b) autonomous vehicles, highlighting points of friction across ethical principles. Possible ways forward towards the implementation of governance on AI are finally examined.

Three companies – Samsung, IBM and Tencent – dominate the global AI patent race over the past 10 years, while fierce competition between the U.S, and China overshadows other countries and regions, including the EU. These are the key findings of OxFirst, a specialist in IP law and economics (and spin out of Oxford University), which also reported that multiple neural nets, machine learning and speech recognition are driving the market. "Patents are mainly filed in the area of interconnectivity and system architecture, suggesting that top players focus primarily on protecting technologies covering multiple neural nets," OxFirst said in its announcement today. "Other areas of crucial importance are ML and bootstrap methods, alongside procedures used during speech recognition processes; e.g. the further establishment of human-machine dialogue." OxFirst said its sector-specific analysis suggests that major companies have focused on AI in the medical space, particularly medical diagnosis, medical simulation and data mining.

Since widespread protests over racial inequality began, IBM announced it would cancel its facial recognition programs to advance racial equity in law enforcement. Amazon suspended police use of its Rekognition software for one year to "put in place stronger regulations to govern the ethical use of facial recognition technology." But we need more than regulatory change; the entire field of artificial intelligence (AI) must mature out of the computer science lab and accept the embrace of the entire community. We can develop amazing AI that works in the world in largely unbiased ways. But to accomplish this, AI can't be just a subfield of computer science (CS) and computer engineering (CE), like it is right now.

The web accessibility market has undergone a tremendous amount of upheaval over the past five years. Most recently, the societal aftershocks of the coronavirus pandemic have reminded everyone of the importance of universal access to digital services. Since 2015, there has also been an explosion of litigation, including class-action lawsuits filed under the ADA (Americans with Disabilities Act) against organizations that have failed to make their websites accessible to disabled people. In 2018, the number of web accessibility lawsuits in the U.S. increased by 177% from the previous year to 2,258. Up to 20% of the population have a disability, be it visual, auditory, or motor, requiring a computer access intervention.