The tech industry is attempting to automate art, with Silicon Valley innovators believing that consumers cannot distinguish between human-created and algorithm-generated content. This week, an AI music platform called Udio beta launched, claiming to allow anyone to create extraordinary music instantly. However, the platform's output has been criticized as acoustical nonsense, with most songs being less than a minute long and of questionable quality.

Despite the platform currently being over capacity for creating new tracks, a review of the existing content revealed a range of unusual and arguably poor-quality tracks. Suggestions have been made that AI music might be best suited for creating simple jingles for advertisements. The article concludes with a critique of the platform's attempt at automated folk music and a sarcastic suggestion that country singer Garth Brooks might consider covering one of the AI-generated songs.

👉 https://gizmodo.com/dune-the-broadway-musical-and-8-other-brain-dead-songs-1851403849

The Israel Defense Forces (IDF) have been using artificial intelligence (AI) systems for identifying targets in Gaza, according to an investigation by journalist Yuval Abraham. The AI system, known as Habsora, identifies "buildings and structures that the army claims militants operate from." Another system, Lavender, is used to compile a "kill list" of suspected combatants. The use of AI in warfare is not limited to Israel, with at least 30 countries now operating defense systems that have autonomous modes. The Pentagon is also developing a response to AI-powered drone swarms, a threat that has become less distant as it counters drone attacks by Yemeni Houthis in the Red Sea.

The use of AI in warfare raises questions about accountability for attacks conducted with little or no human input or oversight. The destruction of Gaza, which resulted in the killing of more than 30,000 Palestinians, a majority of them civilians including more than 13,000 children, offers a vision of war waged by AI. The IDF has denied using an AI system that identifies terrorist operatives, stating that "information systems are merely tools for analysts in the target identification process." However, the IDF has previously bragged about its use of AI in targeting Hamas.

The future of AI in warfare is uncertain, with some experts suggesting that autonomous swarms could battle each other independently, or that AI could be given authority over existing nuclear arsenals. Paul Scharre of the Center for a New American Security proposes a five-point plan to regulate the use of AI in warfare, including human supervision of military AI, a ban on autonomous weapons that target people, and restrictions on control over nuclear weapons. Without such limits, Scharre warns, humanity risks a future of dangerous, machine-driven warfare.

👉 https://www.nytimes.com/2024/04/10/opinion/war-ai-israel-gaza-ukraine.html

Mistral, a French AI start-up, is being backed by European leaders as a potential rival to OpenAI and Google. The company, founded by Arthur Mensch, has gained significant attention just a year after its establishment. The French government views Mistral as a key player in the AI revolution and has been lobbying for EU policymakers to support the firm. The company's generative AI technology, which rivals that of OpenAI, allows businesses to launch AI-driven products such as chatbots and search functions. Major European firms like Renault and BNP Paribas are already using Mistral's technology.

Despite the support, there are concerns about Mistral's ability to compete with larger American and Chinese competitors and the high costs associated with building a successful AI company. However, the company has managed to raise approximately $540 million and has received backing from Microsoft, Salesforce, and Nvidia. Unlike its competitors, Mistral believes in open-source AI software, arguing that it will make systems safer and fuel economic growth.

Mensch, the CEO of Mistral, has expressed concerns about the dominance of American AI giants and their potential impact on global cultures. He has been advocating for lighter regulation, warning that restrictions could hinder innovation. Recently, France successfully lobbied to limit regulation of open-source AI systems in the EU's new Artificial Intelligence Act, a move that benefits Mistral.

👉 https://www.nytimes.com/2024/04/12/business/artificial-intelligence-mistral-france-europe.html

Priyal Shah, a journalism scholar pursuing the Erasmus Mundus Journalism Master's degree, has previously worked as a reporter in India. Her research interests include press freedom and comparative journalism research. The news media industry is currently undergoing a significant transformation, largely due to the Grand Restructuring of Advertising Revenue. This shift has left many publications vulnerable, with effects being felt globally.

👉 https://thefix.media/2024/04/11/ai-in-newsrooms-three-interesting-use-cases/

The media management industry is undergoing a significant transformation due to the Grand Restructuring of Advertising Revenue. This shift has left many publications vulnerable and its impact is being felt globally. The Fix, a one-stop shop for publishers, offers insights, solutions, and data to help navigate this complex landscape.

👉 https://thefix.media/2024/04/03/nyt-voicing-articles-with-ai-assault-on-a-journalist-in-london-weekly-news-digest/

Machine-learning technologies are being used in film restoration for new home video releases, but the results have sparked controversy among viewers. The technology allows for the removal of imperfections in the original film, resulting in a clearer and sharper image. However, some viewers argue that this process removes the authenticity of the film and creates an unnatural appearance. Films like "Titanic," "The Abyss," "True Lies," and "Aliens" have been released in newly restored versions, using artificial intelligence for restoration. The New Zealand company, Park Road Post Production, used proprietary machine-learning software to clean up these films, resulting in a level of detail that some viewers find uncanny.

The use of AI in film restoration has been met with criticism, with some viewers arguing that it is unnecessary for films that already looked fine. The recent releases of James Cameron's films have been the subject of intense scrutiny and debate online, with many describing the images as overly sanitized and artificial. Despite the criticism, the general manager at Park Road Post argues that technology is changing and films are being enhanced for new viewing platforms with higher resolutions. The AI used in film restoration is different from generative AI, which creates new information. AI upscaling, used in film restoration, refines an image without inventing new material. However, the use of AI in film restoration has been stigmatized due to misunderstandings about the technology.

👉 https://www.nytimes.com/2024/04/13/movies/ai-blu-ray-true-lies.html

The podcast "Hard Fork" by Kevin Roose and Casey Newton from The New York Times and Platformer respectively, discusses the impact of AI on jobs and the economy. They explore how companies like Walmart, Wendy's, Abercrombie and Fitch are using AI to boost productivity and potentially reduce their workforce. However, the data on the impact of AI on job loss is still inconclusive.

The hosts also discuss the use of AI in filmmaking with artist Paul Trillo, who used OpenAI’s Sora tool. They also touch on a cyber sneak attack that could have potentially brought down the web but was caught in time. The podcast ends with a discussion on the reluctance of some corporate leaders to replace workers with AI, even when the technology is capable, due to the value of human learning and development in the job.

👉 https://www.nytimes.com/2024/04/05/podcasts/hardfork-ai-jobs-sora-xz.html

Zach Seward, the editorial director of AI initiatives at The New York Times, recently gave a talk at SXSW about the current state of AI-powered journalism. He discussed both the failures and successes of AI in journalism, highlighting the importance of rigorous vetting, reader-focused motivation, and adherence to journalistic principles of truth and transparency. Seward criticized the misuse of AI in generating inaccurate and plagiarized content, as seen in examples from CNET and G/O Media.

On a positive note, Seward also highlighted successful uses of AI in journalism. He mentioned how Quartz used AI to analyze a large cache of documents related to offshore wealth, and how The Texas Observer and Grist used machine learning to identify patterns in data about abandoned oil wells. Other examples included BuzzFeed News using AI to identify surveillance aircraft patterns, The Wall Street Journal using image recognition to detect lead cabling, and The New York Times using satellite imagery for investigations. Seward emphasized that AI can be a powerful tool in journalism when used responsibly and transparently.

👉 https://www.niemanlab.org/2024/03/ai-news-thats-fit-to-print-the-new-york-times-editorial-ai-director-on-the-current-state-of-ai-powered-journalism/

Spotify has introduced a new feature called "daylist" that provides users with three personalized playlists each day, each with a unique and often quirky title. These titles, which are generated by AI, have caught the attention of social media users, boosting the service's popularity. The titles are created using machine learning, which pulls together thousands of descriptors to create unique playlist names. The team behind these playlists includes data scientists and music experts who identify musical descriptors based on genre, mood, and themes.

The daylist feature builds on the popularity of Spotify Wrapped, a year-end review of a user's personalized listening history. The feature also aligns with Spotify's broader strategies around hyper-specificity, with four out of five Spotify users pointing to the platform's personalized offerings as what they like most about the brand. However, the playful brand voice can be a double-edged sword, as corporations risk running afoul of consumer sensitivities with each cheeky ad or brazen tweet. Despite this, users have generally been amused by the daylist feature, with many sharing their unique playlist titles on social media.

👉 https://www.nytimes.com/2024/01/24/style/ai-spotify-music-playlist-algorithm.html

The debate over the intellectual property rights of content generated by artificial intelligence (AI) systems has been a hot topic in the media and entertainment industries. Despite recent legal battles, the author argues that AI is unlikely to pose a significant threat to the livelihoods of creative professionals. Drawing on historical examples, the author suggests that technologies that have made art easier to produce have not stifled human creativity. AI-generated content, while impressive, often lacks the depth and nuance of human-created work.

The author also discusses the legal complexities surrounding AI and intellectual property. AI systems are often trained using large amounts of copyrighted content, raising questions about whether artists should be compensated for their contributions. The author argues against this, likening AI's learning process to a human drawing inspiration from existing work. However, the potential for AI to create near-exact copies of existing works could lead to copyright infringement issues. The author suggests that while AI should not be allowed to create direct copies, it should be given the same freedom to remix art as humans.

👉 https://www.nytimes.com/2023/08/25/opinion/ai-art-intellectual-property.html

The article discusses the challenges in measuring the intelligence and capabilities of artificial intelligence (AI) systems. Currently, there is no standardized method for evaluating AI tools like ChatGPT, Gemini, and Claude, leading to uncertainty about their actual capabilities. This lack of reliable measurement and evaluation is a significant problem as it makes it difficult for users to understand what to do with these AI products. The article also highlights the safety risks associated with poor measurement of AI systems, as it's hard to identify which products might pose real threats of harm.

The most common test for AI models today is the Massive Multitask Language Understanding (MMLU), a collection of multiple-choice questions covering various academic subjects. However, the effectiveness of this test is questionable, and it may soon need to be replaced by more challenging evaluations. The article suggests that the solution to this problem could be a combination of public and private efforts, including robust testing programs by governments, research projects for new evaluations, and commitments from AI companies to work with third-party evaluators.

👉 https://www.nytimes.com/2024/04/15/technology/ai-models-measurement.html

After nearly two years of relying on government grants, Tenomix co-founder and CEO Saumik Biswas decided in January 2023 to seek venture capital for the company's next growth phase. However, the timing was not ideal due to the challenging market conditions, including rising interest rates and inflation that had burst the VC bubble inflated during the pandemic.

Despite these challenges, Biswas was determined to raise funds and began pitching to investors. Many investors, however, were waiting out the challenging market before investing in new startups.

👉 https://thelogic.co/news/special-report/vc-deal-making-plummets-as-ai-hype-softens/

The Financial Times (FT) has been consistent in its strategic and editorial priorities, focusing on subscriptions and digital revenue. The FT Group CEO, John Ridding, has been at the helm since 2006 and has overseen the company's transition to digital. The FT, owned by Japanese company Nikkei, Inc. since 2015, has adapted to changing times, rebranding its magazine "How To Spend It" to HTSI and its enterprise division to FT Professional. The company has also invested in data science, paywall innovation, and consulting services. Currently, the FT has over 1.4 million subscribers, including more than 1 million digital subscribers, and aims to reach a global paying audience of 3 million by 2028.

Ridding believes that the FT's success is due to its commitment to editorial independence and excellence, focus on its core subscription business, and its ownership model. Unlike public ownership and stock market listings, the FT's private ownership by a staff-owned business allows for long-term investment and growth. Despite industry disruptions and challenges, the FT has seen significant growth in its revenues and paid circulation since Nikkei's acquisition. Ridding also highlighted the importance of reader engagement, which the FT measures using a metric called "RFV" (recency, frequency, volume). The company has seen double-digit growth in engaged readership over the past five years.

Ridding sees significant growth opportunities for the subscription and engagement model, particularly through technological advancements like AI and geographical expansion. The FT has been expanding its subscriber base in the U.S. and India, focusing on deepening engagement with its readers. The company is also exploring new storytelling formats and personalization techniques to drive reader engagement.

👉 https://www.niemanlab.org/2024/03/dont-expect-help-from-the-disruptors-the-fts-chief-executive-on-ai-loyalist-readers-and-its-u-s-expansion/

Microsoft has announced a $1.5 billion investment in G42, an Emirati artificial intelligence company with ties to China. The deal, largely orchestrated by the Biden administration, aims to counter China's technological influence in the Gulf region. As part of the agreement, Microsoft will allow G42 to sell its services that use powerful AI chips, while G42 will use Microsoft’s cloud services and comply with a security arrangement negotiated with the U.S. government. This includes protections on the AI products shared with G42 and an agreement to remove Chinese equipment from G42’s operations.

The investment is seen as a strategic move by the U.S. to protect its intellectual property in AI programs and push back against China's rising influence in the Gulf region. However, the deal is sensitive due to G42's ties to China, which include partnerships with Chinese firms and employees from government-connected companies. Despite these concerns, the U.S. believes the agreement ensures that AI technologies can be safely developed, protected, and deployed. The deal could also serve as a model for how U.S. firms can use their technological leadership in AI to attract countries away from Chinese tech, while also reaping significant financial rewards.

The agreement also stipulates that G42 will stop using Huawei telecom equipment, which the U.S. fears could provide a backdoor for Chinese intelligence agencies. G42 is also committed to seeking permission before sharing its technologies with other governments or militaries and is prohibited from using the technology for surveillance. Microsoft will have the power to audit G42’s use of its technology. Despite these measures, concerns remain about less overt ties between G42 and China.

👉 https://www.nytimes.com/2024/04/16/technology/microsoft-g42-uae-ai.html

Microsoft is investing $1.5 billion in G42, an Emirati artificial intelligence company, in a deal facilitated by the White House. The move is seen as an attempt to limit Beijing's tech influence in the Gulf region. As part of the agreement, G42 will sell Microsoft services that use powerful AI chips and will use Microsoft’s Azure cloud services for its AI offerings. G42 has also agreed to remove equipment from Chinese companies like Huawei from its systems, addressing U.S. concerns about potential backdoors for Chinese intelligence agencies.

In other news, Microsoft's president, Brad Smith, will join G42's board, and the company will be able to audit G42’s use of its technology. The deal is expected to benefit Microsoft by giving it a stake in another promising AI company and a foothold to reach more customers, particularly in Middle Eastern countries eager to invest in AI. However, some of these investments are being scrutinized by Washington antitrust regulators.

In political news, big-name donors have poured millions into key races nationwide as the race for the White House tightens and control of Congress remains up for grabs. Democrats started the year with a fund-raising lead, but in the past quarter, Republican campaigns and causes have received significant donations from wealthy backers. The Biden administration's antitrust enforcers are reportedly preparing their next big fight against Live Nation, the parent company of Ticketmaster, following a ticket debacle involving Taylor Swift in 2022.

👉 https://www.nytimes.com/2024/04/16/business/dealbook/washington-microsoft-ai-deal.html

The media management industry is undergoing a significant transformation due to the Grand Restructuring of Advertising Revenue. This shift has left many publications vulnerable and its impact is being felt globally. The Fix, a one-stop shop for publishers, offers insights, solutions, and data to help navigate this complex landscape.

👉 https://thefix.media/2024/04/16/can-ai-help-build-a-pan-european-news-publisher/

The rapid development of generative AI technology is challenging traditional concepts of creativity, authorship, and intellectual property, causing legal complications in Hollywood and beyond. The technology's ability to generate new content based on vast amounts of input data has led to numerous lawsuits over unlicensed use of copyrighted works to train AI models. The legal question of whether this process constitutes actual reproduction of the works is yet to be resolved, with AI companies claiming fair use and plaintiffs struggling to prove actual harm from the use of their works.

The rise of deepfakes, which use AI to mimic individuals without their consent, has also raised legal concerns. Despite the introduction of bills to regulate the use of deepfakes and establish a new class of federal intellectual property, the legality of AI deepfakes remains unclear. The Federal Trade Commission is considering extending its rule against impersonation to cover individuals, while some in the industry are promoting greater transparency in AI training data.

👉 https://variety.com/2024/biz/spotlight/hollywood-ai-concerns-legal-challenges-1235973275/

Microsoft is investing $1.5 billion in UAE-based AI firm G42, acquiring a minority stake and a seat on the board. The partnership will see G42 running its AI applications and services on Microsoft's Azure cloud computing platform, aiming to deliver advanced AI solutions to global public sector clients and large enterprises. The companies will work together to bring advanced AI and digital infrastructure to countries in the Middle East, Central Asia, and Africa. As part of the deal, G42 will cease using Huawei telecom equipment amid US concerns over potential backdoors for Chinese intelligence agencies.

In other news, Google has fired 28 employees involved in a sit-in protest over a $1.2 billion contract with Israel. The company has been under scrutiny for its handling of employee protests and this latest move is likely to fuel further controversy. Meanwhile, Prince Harry has officially renounced his British residency, listing the US as his 'new country'. The former royal has been living in the US with his wife Meghan Markle and their children since stepping back from royal duties in 2020.

👉 https://nypost.com/2024/04/16/tech/microsoft-to-invest-1-5-billion-in-emirati-ai-firm-g42-takes-minority-stake/

A new mobile app called AngryGF has been launched, aiming to teach communication skills by simulating arguments with a virtual girlfriend. The app, available on both the Apple App Store and Google Play, offers various scenarios where the user has to appease their angry AI partner. The user has 10 attempts to say soothing things to increase the "forgiveness level" from 0 to 100 percent. However, the app has been criticized for its irritating and antagonizing responses, leading to questions about its effectiveness and purpose.

The app was created by Emilia Aviles, a Chicago-based social media marketer, who was inspired by her own past relationships. The app is built on OpenAI’s GPT-4 and does not use any additional custom training data. Despite its unique approach, the app has raised concerns about its potential to reinforce negative stereotypes and its lack of consultation with relationship therapists during its development.

👉 https://www.wired.com/story/what-if-your-ai-girlfriend-hated-you/

The U.S. Air Force and Defense Advanced Research Projects Agency (DARPA) have successfully tested a new artificial intelligence system using the experimental X-62A VISTA aircraft. The tests involved AI dogfights between the X-62A and a human-piloted F-16 aircraft, marking the first use of machine-learning-based autonomy in flight-critical systems. Despite previous bans on machine learning tools due to high risk and lack of independent control, the tests were conducted without the need for human pilots to disengage the AI.

The X-62A VISTA is not only a platform for research and advancing the state of tests, but also for preparing the next generation of test leaders, according to Col. James Valpiani, the commandant of the Test Pilot School. The U.S. military began experimenting with AI systems in aircraft in the 1980s under the Strategic Computing Initiative (SCI). The current advancements in machine learning could potentially revolutionize this field. However, Bill Gray, the head of the U.S. Air Force Test Pilot School, clarified that the end goal is not dogfighting, but testing autonomous AI systems in the air for a wider range of applications.

👉 https://gizmodo.com/darpa-tests-ai-dogfighting-computer-skynet-x-62a-vista-1851419381