Do you want to understand why your competitors are winning the business of potential clients? How do you predict future trends within your marketplace? Are you looking to predict government policy decisions that will affect you and your business? These are the questions that can be answered by a team of data scientists that will improve your business and your understanding of your clients. Let's take the first question: "Why are your competitors winning the business of potential clients that you may be missing out on."

A screen shows a demonstration of SenseTime Group's SenseVideo pedestrian and vehicle recognition system at the company's showroom in Beijing. Facial recognition supporters in the US often argue that the surveillance technology is reserved for the greatest risks -- to help deal with violent crimes, terrorist threats and human trafficking. And while it's still often used for petty crimes like shoplifting, stealing $12 worth of goods or selling $50 worth of drugs, its use in the US still looks tame compared with how widely deployed facial recognition has been in China. A database leak in 2019 gave a glimpse of how pervasive China's surveillance tools are -- with more than 6.8 million records from a single day, taken from cameras positioned around hotels, parks, tourism spots and mosques, logging details on people as young as 9 days old. The Chinese government is accused of using facial recognition to commit atrocities against Uyghur Muslims, relying on the technology to carry out "the largest mass incarceration of a minority population in the world today."

A screen shows a demonstration of SenseTime Group's SenseVideo pedestrian and vehicle recognition system at the company's showroom in Beijing. Facial recognition supporters in the US often argue that the surveillance technology is reserved for the greatest risks -- to help deal with violent crimes, terrorist threats and human trafficking. And while it's still often used for petty crimes like shoplifting, stealing $12 worth of goods or selling $50 worth of drugs, its use in the US still looks tame compared with how widely deployed facial recognition has been in China. A database leak in 2019 gave a glimpse of how pervasive China's surveillance tools are -- with more than 6.8 million records from a single day, taken from cameras positioned around hotels, parks, tourism spots and mosques, logging details on people as young as 9 days old. The Chinese government is accused of using facial recognition to commit atrocities against Uyghur Muslims, relying on the technology to carry out "the largest mass incarceration of a minority population in the world today."

The UK government has been funneling millions of dollars into a prediction tool for violent crime that uses artificial intelligence. Now, officials are finally ready to admit that it has one big flaw: It's completely unusable. Police have already stopped developing the system called "Most Serious Violence" (MSV), part of the UK's National Data Analytics Solution (NDAS) project, and luckily was never actually put to use -- yet plenty of questions about the system remain. The tool worked by assigning people scores based on how likely they were to commit a gun or knife crime within the next two years. Two databases from two different UK police departments were used to train the system, including crime and custody records.

Human rights organization Liberty is claiming a win in its native Britain after a court ruled that police trials of facial recognition technology violated privacy laws. The Court of Appeal ruled that the use of automatic facial recognition systems unfairly impacted claimant Ed Bridges' right to a private life. Judges added that there were issues around how people's personal data was being processed, and said that the trials should be halted for now. The court also found that the South Wales Police (SWP) had not done enough to satisfy itself that facial recognition technology was not unbiased. A spokesperson for SWP told the BBC that it would not be appealing the judgment, but Chief Constable Matt Jukes said that the force will find a way to "work with" the judgment.

The world must come together to take action on killer robots, according to a new report. There is increasing agreement among various countries that fully autonomous weapons should be banned to avoid the creation of such killer robots, the new report warns. It would be "unacceptable" if weapons systems are able to select and kill targets without human oversight, the researchers warn. The research by Human Rights Watch said 30 countries had now expressed a desire for an international treaty introduced which says human control must be retained over the use of force. The new report, "Stopping Killer Robots: Country Positions on Banning Fully Autonomous Weapons and Retaining Human Control", reviews the policies of 97 countries that have publicly discussed killer robots since 2013.

AI could also have a transformative effect on clinical decision-making through the utilisation of the huge levels of genomic, biomarker, phenotype, behavioural, biographical and clinical data that is generated across the health system. Bayer and Merck & Co provide a perfect example of this. They have developed an AI software system to support clinical decision-making of chronic thromboembolic pulmonary hypertension (CTEPH) – a rare form of pulmonary hypertension. The software helps differentiate patients from those suffering with similar symptoms that are actually a result of asthma and chronic obstructive pulmonary disease (COPD), and therefore diagnose CTEPH more reliably and efficiently. The CTEPH Pattern Recognition Artificial Intelligence obtained FDA Breakthrough Device Designation in December 2018.

An AI-controlled fighter jet will battle a US Air Force pilot in a simulated dogfight next week -- and you can watch the action online. The clash is the culmination of DARPA's AlphaDogfight competition, which the Pentagon's "mad science" wing launched to increase trust in AI-assisted combat. DARPA hopes this will raise support for using algorithms in simpler aerial operations, so pilots can focus on more challenging tasks, such as organizing teams of unmanned aircraft across the battlespace. The three-day event was scheduled to take place in-person in Las Vegas from August 18-20, but the COVID-19 pandemic led DARPA to move the event online. Before the teams take on the Air Force on August 20, the eight finalists will test their algorithms against five enemy AIs developed by Johns Hopkins Applied Physics Laboratory.

Any technology is a double aged sword. It has Pros and cons. And while pros are imperative in governing any organization, the cons reflect the flaws of technology which is hard to ignore. Hence Artificial Intelligence is no exception for being a threat, especially for cybersecurity. Over the years, the threat regarding AI concerning cybersecurity has grown.

Robustness to distribution shift has been the focus of a long line of work in machine learning [SG86; WK93; KHA99; Shi00; SKM07; Qui 09; Mor 12; SK12]. At a high-level, the goal is to ensure that models perform well not only on unseen samples from the datasets they are trained on, but also on the diverse set of inputs they are likely to encounter in the real world. However, building benchmarks for evaluating such robustness is challenging--it requires modeling realistic data variations in a way that is well-defined, controllable, and easy to simulate. Prior work in this context has focused on building benchmarks that capture distribution shifts caused by natural or adversarial input corruptions [Sze 14; FF15; FMF16; Eng 19a; For 19; HD19; Kan 19], differences in data sources [Sae 10; TE11; Kho 12; TT14; Rec 19], and changes in the frequencies of data subpopulations [Ore 19; Sag 20]. While each of these approaches captures a different source of real-world distribution shift, we cannot expect any single benchmark to be comprehensive. Thus, to obtain a holistic understanding of model robustness, we need to keep expanding our testbed to encompass more natural modes of variation.