Artificial Intelligence Technology Solutions, Inc. along with its wholly owned subsidiary Robotic Assistance Devices, Inc. (RAD) announced revisions to its plans and positioning for a robotic dog for the security services, logistics, utilities, and property management industries. "We're seeing a larger and more cost-conscious market emerge for a RAD dog," said Steve Reinharz, CEO of AITX and RAD. "These market conditions have prompted our team to revise plans and develop a new dog, one that we are naming'CASSIE'. We are now going to bring an entry-level priced'junkyard dog' to market, loaded with all the AI power that RAD has developed over the years." CASSIE (Crawling Autonomous Safety Security Inspection Equipment) is the latest officially announced addition to the RAD family of robotic security and safety solutions and marks the Company's second mobile robot.

Numerous complex questions remain for regulators, lawyers, programmers, creatives and inventors, particularly in light of the rapid, continual rise of AI.

Disturbing footage emerged this week of a chess-playing robot breaking the finger of a seven-year-old child during a tournament in Russia. Public commentary on this event highlights some concern in the community about the increasing use of robots in our society. Some people joked on social media that the robot was a "sore loser" and had a "bad temper". Of course, robots cannot actually express real human characteristics such as anger (at least, not yet). But these comments do demonstrate increasing concern in the community about the "humanisation" of robots.

We propose a fairness measure relaxing the equality conditions in the popular equal odds fairness regime for classification. We design an iterative, model-agnostic, grid-based heuristic that calibrates the outcomes per sensitive attribute value to conform to the measure. The heuristic is designed to handle high arity attribute values and performs a per attribute sanitization of outcomes across different protected attribute values. We also extend our heuristic for multiple attributes. Highlighting our motivating application, fraud detection, we show that the proposed heuristic is able to achieve fairness across multiple values of a single protected attribute, multiple protected attributes. When compared to current fairness techniques, that focus on two groups, we achieve comparable performance across several public data sets.

Machine learning (ML) is increasingly being used to support high-stakes decisions, a trend owed in part to its promise of superior predictive power relative to human assessment. However, there is frequently a gap between decision objectives and what is captured in the observed outcomes used as labels to train ML models. As a result, machine learning models may fail to capture important dimensions of decision criteria, hampering their utility for decision support. In this work, we explore the use of historical expert decisions as a rich -- yet imperfect -- source of information that is commonly available in organizational information systems, and show that it can be leveraged to bridge the gap between decision objectives and algorithm objectives. We consider the problem of estimating expert consistency indirectly when each case in the data is assessed by a single expert, and propose influence function-based methodology as a solution to this problem. We then incorporate the estimated expert consistency into a predictive model through a training-time label amalgamation approach. This approach allows ML models to learn from experts when there is inferred expert consistency, and from observed labels otherwise. We also propose alternative ways of leveraging inferred consistency via hybrid and deferral models. In our empirical evaluation, focused on the context of child maltreatment hotline screenings, we show that (1) there are high-risk cases whose risk is considered by the experts but not wholly captured in the target labels used to train a deployed model, and (2) the proposed approach significantly improves precision for these cases.

Perceptual hashes map images with identical semantic content to the same $n$-bit hash value, while mapping semantically-different images to different hashes. These algorithms carry important applications in cybersecurity such as copyright infringement detection, content fingerprinting, and surveillance. Apple's NeuralHash is one such system that aims to detect the presence of illegal content on users' devices without compromising consumer privacy. We make the surprising discovery that NeuralHash is approximately linear, which inspires the development of novel black-box attacks that can (i) evade detection of "illegal" images, (ii) generate near-collisions, and (iii) leak information about hashed images, all without access to model parameters. These vulnerabilities pose serious threats to NeuralHash's security goals; to address them, we propose a simple fix using classical cryptographic standards.

The corporate landscape of today is frequently in motion. Companies are confronted to uphold their enterprises appropriate in the face of fluctuating markets, whether these fluctuations are social, financial, or mechanical. No matter how successful a company is, it is susceptible to current trends and constantly embryonic new technology. Fortunately, some corporate executives are able to adjust to changing circumstances. Some people are not only adaptable, but they can also take advantage of changing habits and game-changing new technology to achieve massive growth. Technology has correspondingly upgraded our capability to communicate with one another. The rise of mobile technology has blended almost seamlessly with communication software to create a hyper-real web of real-time information, whether it's having your co-workers and employees available via text or video chat at any time or being able to send targeted promotional email blasts to prequalified customers while they're shopping at nearby businesses.

In the years to come, artificial intelligence will be a key feature of cross border criminal investigations, a joint report by Eurojust and eu-LISA, the union's official IT agency found. AI technologies can increase cooperation between EU member states in tackling crime, however, authorities must be careful since machine learning algorithms are prone to biases. AI was listed as a priority in the EU's e-Justice Action plan for 2019-2023. In a world where crime is borderless and criminals employ sophisticated communication tools and technologies, including encryption and AI; tackling crime requires cross-border cooperation by EU Member States and the application of technologies on par with those used by the criminal groups, urged Friday's report. "The field of justice is undergoing digital transformation, and artificial intelligence, as a set of different technologies, has great potential to contribute to and further enhance this process, allowing for a significant improvement in both the efficiency and effectiveness of operation of the judicial authorities," the report said.

A delivery driver who is suing Uber Eats in London over his dismissal from the company and claims its facial recognition technology is racially biased says the company treats couriers as "numbers rather than humans". Pa Edrissa Manjang worked for Uber Eats between November 2019 and April 2021 while employed full-time as a financial assistant. When Manjang first began working for the company he was not regularly asked to send in pictures of himself for verification purposes. However, these facial verification checks became more frequent. Manjang was eventually dismissed from the company by email, when it claimed there were "continued mismatches" between the pictures he took to register for a shift and the one on his Uber work profile.

Artificial Intelligence (AI) has the potential to transform many aspects of life and the UK government has recognized that it is important to review IP laws to ensure that they evolve and promote innovation in this fast-paced area of technology. That was the motivation behind a recent UKIPO consultation which reported earlier this week. With regards to patent protection for AI-devised inventions, the report concluded that no changes are required to UK patent law, at least for the time being. At present, despite claims of certain parties and the international court case relating to the DABUS system which its promotors sought to name as the inventor on patent applications in a number of countries, there is no evidence of AI currently having the capacity to invent. Rather, the general consensus from respondents was that AI technology cannot, at least at present, invent without human assistance.