In a blog post this week, the Federal Trade Commission signaled that it's taking a hard look at bias in AI, warning businesses that selling or using such systems could constitute a violation of federal law. "The FTC Act prohibits unfair or deceptive practices," the post reads. "That would include the sale or use of – for example – racially biased algorithms." The post also notes that biased AI can violate the Fair Credit Reporting Act and the Equal Credit Opportunity Act. "The FCRA comes into play in certain circumstances where an algorithm is used to deny people employment, housing, credit, insurance, or other benefits," it says.

By Nikhil Bedi & Vivek Bhamodkar With evolving business models, increased use of tech and a changing regulatory landscape, fraud management is fraught with newer and more complex challenges. These are further exacerbated during cross-border investigations, where varied levels of standardisation, languages, local laws and regulations, along with specific cultural attributes, bring additional complexities--mandating an investigation methodology standardisation and requiring tools for quick insights. Emerging technology and artificial intelligence (AI) can help make investigations efficient, generate insights, and/or aid reviews. Optimal use of AI demands the knowledge of'possibilities and limitations' of such techniques, either in the form of'special purpose software' or the ability to combine various methods. As significant a role as these tools and technologies may play in the fight against fraud, use of AI, NLP and other technologies come with their own set of challenges.

The European Commission will this week present its proposal on Artificial Intelligence (AI), seen as a step toward a new regulatory framework, promised by Commission President Ursula von der Leyen in her State of the Union, writes Marie-Françoise Gondard-Argenti. Marie-Françoise Gondard-Argenti is a member of the Employers' Group at the European Economic and Social Committee. It is clear that there is no country or company manager in Europe at the moment that does not support the development of a trustworthy and innovative AI ecosystem, which promotes a human-centric approach and that primarily services people, increasing their well-being. There is no company in Europe that does not understand the need to leverage the EU market to spread the EU's approach to AI regulation globally. However, at the moment, the EU lags behind.

Stepping out in public used to make a person largely anonymous. Unless you met someone you knew, nobody would know your identity. Cheap and widely available face recognition software means that's no longer true in some parts of the world. Police in China run face algorithms on public security cameras in real time, providing notifications whenever a person of interest walks by. China provides an extreme example of the possibilities stemming from recent improvements in face recognition technology.

Translation tools from Google and other companies could be contributing to significant misunderstanding of legal terms with conflicting meanings such as "enjoin," according to research due to be presented at an academic workshop. Google's translation software turns an English sentence about a court enjoining violence, or banning it, into one in the Indian language of Kannada that implies the court ordered violence, according to the new study. "Enjoin" can refer to either promoting or restraining an action. Mistranslations also arise with other contronyms, or words with contradictory meanings depending on context, including "all over," "eventual" and "garnish," the paper said. Google said machine translation is "is still just a complement to specialized professional translation" and that it is "continually researching improvements, from better handling ambiguous language, to mitigating bias, to making large quality gains for under-resourced languages."

Like superheroes capable of seeing through obstacles, environmental regulators may soon wield the power of all-seeing eyes that can identify violators anywhere at any time, according to a new Stanford University-led study. The paper, published the week of April 19 in Proceedings of the National Academy of Sciences (PNAS), demonstrates how artificial intelligence combined with satellite imagery can provide a low-cost, scalable method for locating and monitoring otherwise hard-to-regulate industries. "Brick kilns have proliferated across Bangladesh to supply the growing economy with construction materials, which makes it really hard for regulators to keep up with new kilns that are constructed," said co-lead author Nina Brooks, a postdoctoral associate at the University of Minnesota's Institute for Social Research and Data Innovation who did the research while a Ph.D. student at Stanford. While previous research has shown the potential to use machine learning and satellite observations for environmental regulation, most studies have focused on wealthy countries with dependable data on industrial locations and activities. To explore the feasibility in developing countries, the Stanford-led research focused on Bangladesh, where government regulators struggle to locate highly pollutive informal brick kilns, let alone enforce rules.

When Microsoft spends $19.7 billion on a company whose specialties included voice recognition and artificial intelligence (AI) as part of its health sector strategy, you know that AI in the medical field is here to stay. It only makes sense, then, that regulations regarding the technology would not be far behind. Thanks to a leaked document first reported by Politico, we now have our first look at what such regulations might look like in the European Union. The regulation document largely concerns "high-risk" usages of AI. That's not surprising, as the European Commission originally published a whitepaper in February 2020 outlining ideas for regulating such uses of the technology.

More than 70 advocacy groups have called on the Department of Homeland Security to stop using Clearview AI's facial recognition software. In a letter addressed to DHS Secretary Alejandro Mayorkas and Susan Rice, the director of the White House's Domestic Policy Council, the American Civil Liberties Union, Electronic Frontier Foundation, OpenMedia and other organizations argue "the use of Clearview AI by federal immigration authorities has not been subject to sufficient oversight or transparency." The letter points to a recent BuzzFeed News report that found employees from 1,803 government bodies, including police departments and public schools, have been using the software without many of their bosses knowing about it. The company has given out free trials to individual employees at those organizations hoping that they'll advocate for their agency to sign up for it. Besides the lack of oversight, the letter points to issues like racial bias in facial recognition software and the fact Clearview built its database by scraping websites like Facebook, Twitter and YouTube.

Recently, the research on protecting the intellectual properties (IP) of deep neural networks (DNN) has attracted serious concerns. A number of DNN copyright protection methods have been proposed. However, most of the existing watermarking methods focus on verifying the copyright of the model, which do not support the authentication and management of users' fingerprints, thus can not satisfy the requirements of commercial copyright protection. In addition, the query modification attack which was proposed recently can invalidate most of the existing backdoor-based watermarking methods. To address these challenges, in this paper, we propose a method to protect the intellectual properties of DNN models by using an additional class and steganographic images. Specifically, we use a set of watermark key samples to embed an additional class into the DNN, so that the watermarked DNN will classify the watermark key sample as the predefined additional class in the copyright verification stage. We adopt the least significant bit (LSB) image steganography to embed users' fingerprints into watermark key images. Each user will be assigned with a unique fingerprint image so that the user's identity can be authenticated later. Experimental results demonstrate that, the proposed method can protect the copyright of DNN models effectively. On Fashion-MNIST and CIFAR-10 datasets, the proposed method can obtain 100% watermark accuracy and 100% fingerprint authentication success rate. In addition, the proposed method is demonstrated to be robust to the model fine-tuning attack, model pruning attack, and the query modification attack. Compared with three existing watermarking methods (the logo-based, noise-based, and adversarial frontier stitching watermarking methods), the proposed method has better performance on watermark accuracy and robustness against the query modification attack.

The growing quantity and complexity of data pose challenges for humans to consume information and respond in a timely manner. For businesses in domains with rapidly changing rules and regulations, failure to identify changes can be costly. In contrast to expert analysis or the development of domain-specific ontology and taxonomies, we use a task-based approach for fulfilling specific information needs within a new domain. Specifically, we propose to extract task-based information from incoming instance data. A pipeline constructed of state of the art NLP technologies, including a bi-LSTM-CRF model for entity extraction, attention-based deep Semantic Role Labeling, and an automated verb-based relationship extractor, is used to automatically extract an instance level semantic structure. Each instance is then combined with a larger, domain-specific knowledge graph to produce new and timely insights. Preliminary results, validated manually, show the methodology to be effective for extracting specific information to complete end use-cases.