In debates over the future of artificial intelligence, many experts think of these machine-based systems as coldly logical and objectively rational. But in a new study, Princeton University-based researchers have demonstrated how machines can be reflections of their creators in potentially problematic ways. Common machine-learning programs trained with ordinary human language available online can acquire the cultural biases embedded in the patterns of wording, the researchers reported in the journal Science April 14. These biases range from the morally neutral, such as a preference for flowers over insects, to discriminatory views on race and gender. Identifying and addressing possible biases in machine learning will be critically important as we increasingly turn to computers for processing the natural language humans use to communicate, as in online text searches, image categorization and automated translations.

The Donald Trump administration is planning to expand the use of biometric facial recognition systems at airports around the United States. News of the expansion comes from U.S. Customs and Border Protection's Larry Panetta, who spoke about adoption of facial recognition technology at the Border Security Expo. "We currently have everyone's photo, so we don't need to do any sort of enrollment," Panetta said. "We have access to the Department of State records so we have photos of US Citizens, we have visa photos, we have photos of people when they cross into the US and their biometrics are captured into [Department of Homeland Security database] IDENT." The project to equip airports around the country with facial recognition systems is known as Biometric Exit.

The cyberwarfare landscape is changing -- here's how to prepare The cyberwarfare landscape is changing -- here's how to prepare The cyberwarfare landscape is changing -- here's how to prepare Get the latest federal technology news delivered to your inbox. Machine learning makes life much more manageable for network security operators. If the recent spate of alleged Russian cyberattacks has taught us anything, security breaches can happen so quickly and stealthily, the damage will be done before anyone even realizes there was a hack. In fact, as malicious actors become more insidious, federal network security managers are finding the reaction time between identifying and mitigating potential threats has gone from minutes to milliseconds. Factor in the volume and complexity of the threats, and it becomes evident the challenge has grown well beyond what can be managed through manual intervention.

The EU is launching an unprecedented public consultation today to find out what Europeans fear most about the future of the internet. A succession of surveys over the coming weeks will ask people for their views on everything from privacy and security to artificial intelligence, net neutrality, big data and the impact of the digital world on jobs, health, government and democracy. A dozen leading European publications, including the Guardian, are to publicise the surveys over the coming three weeks. Results will be compiled in early June. Readers can complete the first questionnaire here.

A device that can read people's minds by detecting their brainwaves has been developed in a breakthrough that could eventually enable people with "locked-in syndrome" to communicate. The system was only partially effective with a 90 per cent success rate when trying to recognise numbers from zero to nine and a 61 per cent rate for single syllables in Japanese, the researchers said. But, nonetheless, a statement about the research issued by the Toyohashi University of Technology in Japan said it showed that an effective device to read people's thoughts and relay them to others was possible in the "near future". They even suggested an "easily operated" device with a smartphone app could be ready in just five years. An electroencephalogram (EEG) was used to monitor people's brain waves while they spoke.

Some of the most promising and popular software applications are built in the Pacific Northwest. That much is evident from our nominees for the App of the Year category in the GeekWire Awards, which kicked off online voting today. This year's nominees range from tech giants trying to improve enterprise communication to university researchers trying to revolutionize health screening. Last year's winner was Bellevue-based OfferUp. Finalists for 2017 include Microsoft Teams; Mighty AI; RIPL; Shyft Technologies; and Senosis Health.

Google has been forced to open up Android to rival search engines and applications in Russia, after settling a two-year battle with competition authorities for 439m roubles (£6.2m). The deal sets a new precedent for Google, which until now has resisted permitting the pre-installation of rival search engines and certain applications on to the world's most popular operating system. In 2015, Russia's Federal Antimonopoly Service (FAS) ruled that the Android-maker was breaking the law and abusing its dominant position through restrictions on third-party manufacturers, fining Google 438m rubles (£6.2m) in August 2016. The FAS said that Google will no longer demand exclusivity of its applications on Android devices sold in Russia and will not restrict the pre-installation of rival search engines and other applications. Google will also develop a tool allowing users to choose the default search engine on new and existing Android devices, a similar measure put in place by Microsoft for browser choice on Windows following antitrust action by the European Commission.

Later today, an Atlas V rocket will blast off to the International Space Station bearing 7,600-odd pounds of food, water and other sundries to the astronauts on board. Nearly a third of that weight is equipment for science that furthers NASA's ability to explore space: a capsule that records how spacecraft break up on re-entry, an experiment that tests how fires behave in microgravity, a shiny new LED-powered plant growing habitat. But also tucked into the spacecraft's cargo hold are 38 CubeSats, diminutive satellites that are cheaper and easier to assemble than the behemoths rockets usually take into orbit. The ISS is a major launch platform for CubeSats--it has ejected more than 100 small payloads from companies and universities into orbit in the past few years. Those small sats will rely on new launch providers when the ISS retires in less than a decade. But this launch, from Cape Canaveral Air Force Station in Florida, will be business as usual; you can watch the livestream (now also in glorious 3D!) when it starts at 10 am Eastern, with the launch scheduled for 11:11 am.

AMSTERDAM– On March 25, the European Union celebrated the 60th anniversary of the Treaty of Rome, which gave birth to the European Economic Community and set the stage for the EU's creation. Like any birthday, it was an opportunity not just to celebrate, but also to consider what has been achieved – and what must change. At a time when member countries are increasingly pulling away from the EU – and the United Kingdom is actually leaving – the need for reform is clear. Discussions about how to bring the European Council, Commission, and Parliament closer to EU citizens, thereby closing Europe's "democratic deficit," are frequent and animated. But another institution that could benefit from more input from EU citizens is often left out of these discussions: the European Court of Justice.

The consequences of non-compliance can be crippling for organisations, from losing customer trust to paying hefty fines. Next summer, organisations will need to comply with a new set of EU regulations that aim to ensure good practice when collecting and analysing consumer data. Profit has well and truly shifted from product to digital information – in other words, data is power. The need to protect individual data is becoming more important, especially in light of cybersecurity breaches and targeted advertising. The new European General Data Protection Regulation (GDPR) represents the most important change to data privacy in 20 years, but how can businesses prepare for enforcement, and how will it impact data analysis?