Dear readers,
Welcome to the first Saturday edition of Inside AI! Going forward, this newsletter will deliver to your inbox seven days a week, including PJC partner Rob May's Sunday edition.
Saturday's edition will take a broader approach than our weekday newsletter. You can expect to see in-depth reports, timelines, graphs, and illustrations about emerging AI topics.
Today we'll deliver a recap of the top AI stories (so far) of 2021. Below that, you'll find a survey asking what topic you'd like us to delve into next time.
|
|
|
Top stories of 2021:
- Google launched a "sensitive topics" review of its research papers this year, which led to at least three examples of it asking authors not to portray its technologies in a negative way. The Reuters findings come several weeks after former AI ethics co-lead Timnit Gebru was dismissed after managers asked her to retract an AI research paper or take her name off it. The investigation highlights ongoing concerns about corporate control and its influence on academic research, including in the AI field.
- A New Jersey man who was falsely arrested due to facial recognition has sued local police and the city of Woodbridge. Nijeer Parks is the third person to be wrongfully accused of a crime based on a faulty match from the AI software. All three victims were Black men.
- Skype co-founder Jaan Tallinn listed AI as among the top threats to humanity this century. The Estonian computer programmer told CNBC that he's invested in safety initiatives at universities and AI labs such as Deepmind to make sure the technology is developed ethically and responsibly.
- Forrester predicts the AI software market will be lower than analysts' forecasts over the next five years as AI becomes more mainstream. The market research firm believes more AI technologies will be embedded as a standard feature in software for free, rather than growing as a separate category.
- A Boston Dynamics video featuring its lineup of robots dancing has gone viral. The robotics company's humanoid Atlas, dog-like Spot, and box-moving Handle robots participated in a coordinated dance routine set to The Contours’ song "Do You Love Me."
- A newly discovered patent describes how Microsoft wants to use a person's digital offerings to create a chatbot in their image. Social media profiles, emails, images, and more would be fed into a machine learning system, which would train the bot in order to mimic how that person sounds.
- OpenAI announced two neural networks that can make images out of text, as well as caption images. The ML company introduced the models—respectively named DALL-E and CLIP—in a blog post today. They strive to do for images what OpenAI's GPT-3 has done for text generation, bringing AI closer to better understanding language by "seeing" what words mean.
- A $741 billion defense policy bill passed by Congress greenlights $6.4B worth of federal funding toward AI research and initiatives. Both the Senate and House overrode former President Trump's veto of the legislation, which effectively sets the stage for a U.S. national strategy on AI.
- A New York City councilwoman introduced a proposal that would force companies to disclose their AI-based hiring practices. These range from automated resume screenings to algorithms that analyze job candidates' facial movements and responses in video interviews.
- The White House launched its national AI office, which will oversee the country's AI strategy and coordinate R&D and policies for the federal government. The recently passed National Artificial Intelligence Initiative Act called for the creation of the office, which will also work with the private sector and academia on AI initiatives.
- Stanford researchers claim to have developed an AI algorithm that can predict a person's political orientation based solely on a photo. The algorithm achieved just over 70% accuracy in tests, better than chance (50%) or human accuracy (55%), according to their paper published in Scientific Reports. As The Next Web's Tristan Greene points out, such a study is highly suspicious at the very least, and in his words, "demonstrably false."
- Amazon is opening up its Alexa AI technologies to third-party companies so they can build their own virtual assistants. Through Alexa Custom Assistant, companies like auto manufacturers can customize the voice, features, and wake word of their branded assistants, which will “co-exist” with Alexa, the company said.
- U.S. aviation regulators granted permission to American Robotics to operate fully autonomous drone vehicles for the first time. The approval means the robotics company can fly its smart drones without a human piloting and/or observing them.
- In a first for the AI/ML industry, the parent company of a controversial cloud storage app was forced to delete its algorithms, not just data, OneZero reports. The Federal Trade Commission ordered Paravision, the company behind the Ever cloud storage app, to delete all of its photos taken from customers, as well as algorithms trained on that data.
- Several U.S. lawmakers plan to reintroduce a bill this year that would ban the FBI and other federal agencies from using facial recognition technologies, Reuters reports. The Facial Recognition and Biometric Technology Moratorium Act failed to gain traction last year after it was proposed by four Democratic Congress members in June.
- European Parliament members adopted a new AI report yesterday that calls for a ban on lethal autonomous weapons, as well as "highly intrusive" AI-based social scoring applications. AI programs can't replace human contact or decision-making, according to the report, which provides guidelines for both military and non-military AI uses.
- Google's parent company Alphabet is shutting down Loon, which used a fleet of AI-powered balloons to provide wireless internet connectivity to remote parts of the world. Loon's internet balloons, which began as an Alphabet X moonshot project, have been piloted by a novel, machine-designed deep reinforcement learning system.
- The U.S. maintains a "substantial lead" in artificial intelligence worldwide, though China is closing the gap in some areas, according to a new report by the Information Technology and Innovation Foundation. The researchers predict that China will eventually catch up to the U.S. in its development and use of AI technologies. The European Union, meanwhile, will continue to lag, they predict.
- The World Economic Forum (WEF) launched a global AI alliance that it says is meant to "accelerate the adoption of inclusive, trusted and transparent AI." The alliance already counts more than 100 companies, governments, non-profits, and other organizations as among its ranks.
- Hanson Robotics says it will start mass-producing four robots, including its well-known humanoid bot Sophia, during the first half of 2021. The Hong Kong-based company says more automation is needed during the pandemic to "keep people safe" and provide companionship to ward off loneliness.
|
|
Top stories continued...
- Elon Musk said his company Neuralink has implanted one of its brain-chips into a monkey, allowing the animal to play video games with his mind. The Tesla CEO told Clubhouse, a social media app, that there's no sign of where the wireless neural chip was implanted and the monkey doesn't appear "uncomfortable."
- Ford and Google agreed to a six-year partnership that will see the tech giant integrate its operating systems into the automaker's Ford and Lincoln vehicles, starting in 2023. Ford will use Google's artificial intelligence technology to "improve the efficiency of vehicle development, supply chain and manufacturing operations."
- Boston Dynamics released new tools — including a robotic arm — for its mobile robot Spot. The front-mounted arm can pick up trash, open doors, turn off switches, clean rooms, and more, according to the Hyundai-owned robotics firm.
- U.S. facial recognition startup Clearview AI violated Canada's privacy laws when it gathered photos of people without their knowledge or consent, according to a new report from the country's privacy watchdog. Canada’s Office of the Privacy Commissioner investigated the controversial New York startup, concluding that it conducts illegal mass surveillance.
- Amazon Web Services chief Andy Jassy — who is replacing Jeff Bezos as CEO later this year — has weighed in on AI issues ranging from facial recognition to lethal autonomous weapons. Many reports have noted that Jassy's appointment, announced in February, sends the signal that Amazon's future lies in the cloud and web services, which incorporate AI and machine learning solutions.
- Amazon is installing a driver-monitoring system in some of its delivery vans, which reportedly uses AI and computer vision to detect when drivers are distracted, and other potential dangers. The multi-camera system, supplied by California-based Netradyne, would monitor happenings both inside and outside of the vehicle, ostensibly to improve safety.
- Two Google engineers resigned from the company, citing the recent exit of AI researcher Timnit Gebru. David Baker, an engineering director, and Vinesh Kannan, a software engineer, are the latest Googlers to criticize the tech giant's handling of diversity, ethics, and other issues.
- Ford said it's investing $7B into autonomous vehicles through 2025. The automaker — which rolled out its fourth-generation self-driving test vehicles in October — also announced a $22B investment in electric vehicles during the same time frame.
- Palantir and IBM announced a partnership on open AI applications for businesses using large datasets. The companies will utilize Palantir's operations platform and IBM's hybrid cloud data platform to build an AI tool, "Palantir for IBM Cloud Pak for Data."
- Chinese AI and internet company Baidu is seeking to raise money for a standalone AI chip company, CNBC reports. The search engine giant is in talks with VC firms, including GGV and IDG Capital, to invest in the business, which would manufacture chips for automakers and other industries.
- The Spaceborne Computer-2 (SBC-2) will bring AI and machine learning capabilities to the International Space Station within the next three years. The edge computing system from Hewlett Packard Enterprise will help ISS astronauts process and transmit large amounts of data, speeding up their experiments.
- Nissan announced it isn't in talks with Apple to build an electric autonomous vehicle. The Financial Times reported Apple approached Nissan about a potential partnership to create self-driving cars, but talks fell apart after both sides disagreed on branding. The tech giant wanted the vehicle to be branded as the "Apple Car," however Nissan felt its role would be reduced to that of a hardware supplier.
- A coalition of civil-rights groups sent a letter to President Biden asking him to halt the federal government's use of facial recognition, calling it “extraordinarily dangerous to core freedoms even when it works exactly as advertised.” The 40+ organizations are also asking Biden to block state and local governments from using federal funds to access the AI technology.
- Google is restructuring its responsible AI efforts after the effective firing of AI researcher Timnit Gebru. In a blog post, Google said engineering VP Marian Croak will now oversee "a new center of expertise" at Google Research that's focused on the responsible development of AI technologies.
- Technology being developed to read and alter brain activity could pose privacy risks, warn ethicists, scientists, and Science News readers. According to a survey of readers, privacy was the most worrying aspect of neurotechnology being developed by Elon Musk's Neuralink and other companies, beating out autonomy and fairness by a wide margin.
- NASA's Perseverance robotic rover — which landed on the surface of Mars last month— has the most AI tech and capabilities of any Mars rover yet, according to NASA's Raymond Francis. This includes its AI-enabled Terrain Relative Navigation system, which helped navigate its touchdown, and the Autonomous Exploration for Gathering Increased Science system, allowing engineers to aim and control the rover's chemistry camera.
- IBM is looking to sell its IBM Watson Health business, which generates around $1B annual revenue, the WSJ reports. Watson Health was expected to be one of the company's major moves into AI and healthcare. However, physicians' reluctance to adopt the technology and problems in collecting and analyzing data under medical privacy laws could have led to the division's lack of growth.
- Google says it fired Margaret Mitchell, co-lead of its ethical AI team, due to "multiple violations" of its code of conduct. This comes several weeks after Google opened a review of Mitchell, based on claims that she shared thousands of internal company files with outsiders.
- U.K. spy agency GCHQ uses, or will use, AI to track worldwide criminal networks, fight against cyberattacks, and identify state-backed disinformation campaigns, according to a new report. In "Ethics of AI: Pioneering a New National Security," GCHQ explained how it can apply AI techniques to boost national security and protect the U.K. from outside threats.
- Google plans to alter how it reviews AI ethics papers and work to "regain trust" after losing two prominent AI ethicists, according to a town hall recording obtained by Reuters. Google Research executives discussed ways to improve its reputation for responsible AI after the firings of Margaret Mitchell and Timnit Gebru, who co-led its Ethical AI team.
|
|
Top stories continued...
- Facial-recognition startup Clearview AI will ask the U.S. Supreme Court to clarify Illinois’ Biometric Information Privacy Act (BIPA), a state law that bans the collection of facial recognition images and other data without a person's consent. The company is asking an appeals court to postpone a decision involving a BIPA lawsuit until the high court decides if it wants to take up the case.
- The U.S. is unprepared to compete in the new age of artificial intelligence and risks falling behind China and other superpowers, according to National Security Commission on Artificial Intelligence's latest report. The 756-page report stresses the need for the U.S. to invest in AI-enabled weapons and place "wise restraints" on mass surveillance tools such as facial recognition.
- China surpassed the U.S. in AI research journal citations last year, according to the latest Artificial Intelligence Index report. Stanford University’s Institute for Human-Centered Artificial Intelligence (HAI) released the report, which outlines various benchmarks and statistics about world progress in AI.
- Facebook is looking into facial recognition software for its upcoming smart glasses, hardware chief Andrew “Boz” Bosworth confirmed. In an Instagram Q&A, the executive clarified that dialogue is still needed with the public — given facial recognition's many controversies — and "if people don’t want this technology, we don’t have to supply it."
- McDonald’s is considering a partial sale of its AI company Dynamic Yield Ltd., which developed technology that changes a company's digital menu boards based on weather and other factors. The fast-food chain acquired Dynamic in 2019 for $300M — its largest purchase in two decades — before installing its predictive technology at drive-thrus in the U.S., Canada, and Australia.
- The U.S. government should invest more in R&D funding to accelerate the creation of AI "smart cities," according to a think tank's report. The Information Technology and Innovation Foundation (ITIF) concluded that the federal government should coordinate AI intelligence tools to help U.S. cities jumpstart their smart technology applications.
- The creator of a series of spookily-real Tom Cruise deepfake videos that have gone viral on TikTok says the clips take weeks of work and that the public should not be too concerned with "one-click fakes." Visual effects artist Chris Umé told The Verge he's finished making the videos (after posting a total of four on TikTok) now that he's been able to prove what could be accomplished with the technology. "It's fulfilled its purpose. We had fun," he said.
- China listed AI and brain science as among the “frontier technologies” it plans to focus on in the next five years and beyond. The country is seeking to increase R&D in advanced technologies to boost its domestic offerings and better compete with the U.S., according to its latest five-year development plan.
- The U.S. should raise its R&D budget for AI efforts to $2B in 2022 — up from $1.5B this year — and then double it every year until 2026, argues Eric Schmidt, chair of the U.S. government's National Security Commission on Artificial Intelligence. In remarks shared with CNN's Fareed Zakaria on Sunday, the former Google chairman renewed calls for the U.S. to urgently focus on AI and high tech as matters of national security, ethics, and privacy.
- Civil rights activists and community groups sued Clearview AI in California, alleging that the facial recognition startup collects residents' photos and biometric data without their consent or notice. The plaintiffs are seeking an injunction to halt the company's collection of data in the state and demand the permanent deletion of all data in its databases.
- DeepMind, the London-headquartered AI lab, is reportedly recruiting AI researchers in New York. The research lab, a subsidiary of Google’s parent company Alphabet, hired Facebook AI Research (FAIR) co-founder Rob Fergus last year to oversee the NY team, which is small but will "carefully grow over time," according to a CNBC report.
- Facebook's new Learning from Videos project aims to develop an AI system that can understand the content in videos, part of its broader initiative to build machines "that learn like humans do." The tech giant says it will train its machines on publicly available videos to help improve accuracy.
- Baidu's AI chip business Kunlun has raised an undisclosed amount of funding, valuing the unit at ~$2B. The financing round comes as Baidu weighs bringing its AI chip design to market and making Kunlun a standalone company.
- TikTok owner ByteDance says it's hiring a research team to explore the possible development of AI chips for cloud computing. The startup's foray into chip-making comes as China continues its domestic push into AI and more self-reliance on semiconductors.
- OpenAI's Sam Altman predicts AI technologies will create enough wealth to eventually pay each U.S. adult $13.5K per year. This wealth redistribution from the AI revolution could occur as soon as a decade, according to Altman, who's president and co-founder of the AI nonprofit company co-founded by Elon Musk.
- A New York City Council member wants to ban police agencies from using weaponized robots, after videos of the NYPD's four-legged "Digidog" went viral. Councilman Ben Kallos' bill would not actually outlaw Digidog, a painted version of Boston Dynamics' Spot, which is unarmed.
- The New York Times shared some résumé-writing tips to make it past AI screening software. HR departments are progressively using more AI systems and predictive hiring tools to vet applicants up-front.
- William Shatner is working with tech company StoryFile to create an advanced AI system of himself. The actor, who turned 90 this month, says it will allow viewers to ask the conversational AI program questions about his work and personal experiences.
- Human-like robot Sophia has sold a piece of digital art for nearly $700K. An unidentified buyer acquired a Non-Fungible Token (NFT) that verifies the authenticity of the digital artwork — a 12-second MP4 video file — as well as a physical piece of art painted by Sophia.
|
|
What would you like to see Inside AI focus on?
|
|
|
|
|
|
|
|
|
|
|
|
|
QUICK HITS:
- Monogram is bringing 3D printed custom knee replacements and orthopedics to market. Learn more about how you can invest in this revolutionary company.*
- Apple acquired the most AI companies of any buyer worldwide from 2016-2020, according to GlobalData research, beating out Accenture, Google, Microsoft, and Facebook.
- The European Commission is on track to unveil its first AI regulations next month and some officials are calling for a "CERN for AI."
- Facebook’s F8 developer conference returns this year as a virtual-only event on June 2. The one-day event, F8 Refresh, will be open to developers worldwide and live-streamed on the Facebook for Developers page.
- DARPA is looking into new and more efficient ways to improve computer vision and visually track objects.
- In 2021, AI is expected to generate 80-85 billion roubles ($1.05B-$1.11B) in net profit for Russian lender Sberbank, according to CEO German Gref.
- 100+ companies, including those backed by Google Ventures, Andreessen, Founders Fund, Kleiner and Bloomberg have hired software engineers from this company.*
*This is sponsored content.
|
|
|
|
Beth is a tech writer and former investigative reporter for The Arizona Republic. A graduate of the Walter Cronkite School of Journalism, she won a First Amendment Award and a Pulitzer Prize nomination for reporting on the rising costs of public pensions.
|
|
Editor
|
Charlotte Hayes-Clemens is an editor and writer based in Vancouver. She has dabbled in both the fiction and non-fiction world, having worked at HarperCollins Publishers and more recently as a writing coach for new and self-published authors. Proper semi-colon usage is her hill to die on.
|
|