A une semaine du vote de confiance, Bayrou entame des consultations politiques

François Bayrou entame lundi après-midi une série de consultations des partis politiques, à une semaine du vote de confiance qu’il sollicite de l’Assemblée nationale autour de la question budgétaire, qui pourrait sceller le sort de son gouvernement.Le Premier ministre recevra lundi à 17h00 les représentants du Parti communiste (PCF), son secrétaire national Fabien Roussel et le président du groupe à l’Assemblée, Stéphane Peu.Suivront mardi et mercredi les représentants des partis soutenant la coalition présidentielle, mais aussi ceux de Place publique, le parti de Raphaël Glucksmann, du Rassemblement national, de l’UDR, le parti d’Eric Ciotti, puis du groupe centriste de l’Assemblée Liot.Le Parti socialiste se rendra à Matignon jeudi matin, suivi de l’UDI, présidé par Hervé Marseille.La France insoumise et les Ecologistes ont pour leur part refusé d’honorer cette invitation.Après avoir créé la surprise en annonçant engager la responsabilité de son gouvernement le 8 septembre, avant même le début des discussions budgétaires, François Bayrou apparaît plus qu’en sursis à Matignon.Dimanche, il a de nouveau défendu sa position lors d’un entretien accordé aux quatre chaînes d’information en continu, estimant notamment que la question en jeu lors de ce vote n’était pas “le sort du Premier ministre” mais celui de la France.Il n’y a “aucune politique courageuse possible” sans “accord minimal” sur le “diagnostic”, a également répété dimanche le Premier ministre.Les chances de compromis paraissent minces, M. Bayrou ayant considéré que les propositions budgétaires du PS signifiaient qu'”on ne fait rien” pour réduire l’endettement.Le PS, qui se dit volontaire pour prendre la suite de M. Bayrou à Matignon après sa chute probable, propose notamment de réduire le déficit de 21,7 milliards d’euros en 2026, environ deux fois moins que les 44 milliards visés par le gouvernement.Au programme: 14 milliards d’économies “sans mettre à contribution les travailleurs et les services publics”, et 26,9 milliards de recettes nouvelles, pesant “d’abord sur les grandes fortunes”.”Le plan qu’ils ont sorti cette semaine fait la démonstration qu’ils ne veulent pas gouverner”, a estimé dimanche un ancien ministre macroniste.François Bayrou a présenté le 15 juillet les grandes lignes de son projet de budget comprenant 43,8 milliards d’euros d’effort financier via un certain nombre de mesures –“année blanche fiscale”, gel des prestations sociales– dont la plus commentée est la suppression de deux jours fériés sans contrepartie rémunératrice.Le chef du gouvernement devra, en cas de vote négatif le 8 septembre, présenter la démission de son gouvernement, moins d’un an après avoir succédé à Michel Barnier, renversé par une motion de censure sur les textes budgétaires.

In oil-rich Oman, efforts to preserve frankincense ‘white gold’

The arid Dawkah valley is home to one of Oman’s most prized resources: not oil or gas but frankincense trees, their fragrant sap harvested for millennia by residents who call it “white gold”.Located in Oman’s southern Dhofar region, bordering Yemen, the valley is the world’s largest such reserve, home to around 5,000 frankincense trees that dot the barren earth, their trunks bearing kernels that exude a distinctive woody scent.”For us, frankincense is more precious than gold. It’s a treasure,” said Abdullah Jaddad, a frankincense harvester resting in the shade of a tree.The oil extracted from the sap of the frankincense tree is used in perfume and skincare but it is also sold as solid beads of fragrance in local markets.The high-end Omani perfume-maker Amouage, which manages the reserve, sells its luxury scents internationally for hundreds of dollars a bottle — with one limited edition perfume containing frankincense sold for nearly $2,000.The Dawkah valley is one of the rare places in the world where the Boswellia tree, from which frankincense resin is extracted, grows. Since 2000, it has been a UNESCO World Heritage Site as part of the Land of Frankincense listing, along with Khor Rori, Al Baleed and Shisr.- Like oil -With its unique earthy scent, frankincense has long been used as incense, but also in traditional medicine, and even religious rituals.Before modern technology, the frankincense trade, which began in the third millennium BC, extended from Dhofar via sea and caravan routes to Mesopotamia, the Indus Valley and Ancient Egypt, all the way to Greece, Rome and even China.”Frankincense had roughly the same value as oil today,” according to Ahmed al-Murshidi, who heads the Khor Rori site.The ancient port of Samahram, which forms part of the Khor Rori site, served as the gateway for frankincense to the world.As Jaddad collected dried beads of sap from the trees, he told AFP that the type of frankincense found in the valley was the Najdi — one of four main varieties.The Najdi and Hojari varieties are used for their medicinal properties, according to Faisal Hussein Bin Askar, whose father founded the Bin Askar frankincense shop, in business since the 1950s.”The cleaner and purer the frankincense, the more suitable it is for drinking as a treatment, while the rest is used as incense,” he said, adding that several factories in Dhofar are specialised in frankincense skincare and oils.The highest-grade and rarest frankincense has a light green colour.- ‘Quick to anger’ -The resin is harvested by hand using traditional methods that involve cutting the bark to release the sap and leaving it for a few days to harden.Harvesting the tree requires care and skilled craftsmanship.As one guide put it to a group of tourists at the Land of Frankincense Museum in Salalah: “the frankincense tree is quick to anger”.”We strike the tree in specific, small spots, about five times, to preserve” the plant, said Musallam bin Saeed Jaddad, who works in the reserve.”No one should cut open a frankincense tree… it could kill it,” he said.In 2022, Amouage partnered with Omani authorities to develop the Dawkah reserve and provide jobs for the local community, only harvesting a fifth of the trees to preserve them.Each tree has a unique code and is monitored by a team of specialists, with donations open to anyone wanting to help the reserve in exchange for small gifts of frankincense products every year.A distillery is set to be built in the reserve to extract the frankincense oil, a process for now completed in France, said Mohammed Faraj Istanbuli, the reserve supervisor.”The government is carrying out vital projects, like building roads for example, which threatens other areas where frankincense trees grow,” he said.”We bring those trees… to the reserve. We have saved about 600 trees so far.”

‘Fueling sexism’: AI ‘bikini interview’ videos flood internet

The videos are strikingly lifelike, featuring bikini-clad women conducting street interviews and eliciting lewd comments — but they are entirely fake, generated by AI tools increasingly used to flood social media with sexist content.Such AI slop — mass-produced content created by cheap artificial intelligence tools that turn simple text prompts into hyper-realistic visuals — is frequently drowning out authentic posts and blurring the line between fiction and reality.The trend has spawned a cottage industry of AI influencers churning out large volumes of sexualized clips with minimal effort, often driven by platform incentive programs that financially reward viral content.Hordes of AI clips, laden with locker-room humor, purport to show scantily clad female interviewers on the streets of India or the United Kingdom — sparking concern about the harm such synthetic content may pose to women.AFP’s fact-checkers traced hundreds of such videos on Instagram, many in Hindi, that purportedly show male interviewees casually delivering misogynistic punchlines and sexualized remarks — sometimes even grabbing the women — while crowds of men gawk or laugh in the background.Many videos racked up tens of millions of views — and some further monetized that traction by promoting an adult chat app to “make new female friends.”The fabricated clips were so lifelike that some users in the comments questioned whether the featured women were real.A sample of these videos analyzed by the US cybersecurity firm GetReal Security showed they were created using Google’s Veo 3 AI generator, known for hyper-realistic visuals.- ‘Gendered harm’ -“Misogyny that usually stayed hidden in locker room chats and groups is now being dressed up as AI visuals,” Nirali Bhatia, an India-based cyber psychologist, told AFP.”This is part of AI-mediated gendered harm,” she said, adding that the trend was “fueling sexism.”The trend offers a window into an internet landscape now increasingly swamped with AI-generated memes, videos and images that are competing for attention with — and increasingly eclipsing — authentic content.”AI slop and any type of unlabeled AI-generated content slowly chips away at the little trust that remains in visual content,” GetReal Security’s Emmanuelle Saliba told AFP.The most viral misogynistic content often relies on shock value — including Instagram and TikTok clips that Wired magazine said were generated using Veo 3 and portray Black women as big-footed primates. Videos on one popular TikTok account mockingly list what so-called gold-digging “girls gone wild” would do for money.Women are also fodder for distressing AI-driven clickbait, with AFP’s fact-checkers tracking viral videos of a fake marine trainer named “Jessica Radcliffe” being fatally attacked by an orca during a live show at a water park.The fabricated footage rapidly spread across platforms including TikTok, Facebook and X, sparking global outrage from users who believed the woman was real.- ‘Unreal’ -Last year, Alexios Mantzarlis, director of the Security, Trust, and Safety Initiative at Cornell Tech, found 900 Instagram accounts of likely AI-generated “models” — predominantly female and typically scantily clothed.These thirst traps cumulatively amassed 13 million followers and posted more than 200,000 images, typically monetizing their reach by redirecting their audiences to commercial content-sharing platforms.With AI fakery proliferating online, “the numbers now are undoubtedly much larger,” Mantzarlis told AFP.”Expect more nonsense content leveraging body standards that are not just unrealistic but literally unreal,” he added.Financially incentivized slop is becoming increasingly challenging to police as content creators — including students and stay-at-home parents around the world — turn to AI video production as gig work.Many creators on YouTube and TikTok offer paid courses on how to monetize viral AI-generated material on platforms, many of which have reduced their reliance on human fact-checkers and scaled back content moderation.Some platforms have sought to crack down on accounts promoting slop, with YouTube recently saying that creators of “inauthentic” and “mass produced” content would be ineligible for monetization.”AI doesn’t invent misogyny — it just reflects and amplifies what’s already there,” AI consultant Divyendra Jadoun told AFP.”If audiences reward this kind of content with millions of likes, the algorithms and AI creators will keep producing it. The bigger fight isn’t just technological — it’s social and cultural.”burs-ac/st

‘Fueling sexism’: AI ‘bikini interview’ videos flood internet

The videos are strikingly lifelike, featuring bikini-clad women conducting street interviews and eliciting lewd comments — but they are entirely fake, generated by AI tools increasingly used to flood social media with sexist content.Such AI slop — mass-produced content created by cheap artificial intelligence tools that turn simple text prompts into hyper-realistic visuals — is frequently drowning out authentic posts and blurring the line between fiction and reality.The trend has spawned a cottage industry of AI influencers churning out large volumes of sexualized clips with minimal effort, often driven by platform incentive programs that financially reward viral content.Hordes of AI clips, laden with locker-room humor, purport to show scantily clad female interviewers on the streets of India or the United Kingdom — sparking concern about the harm such synthetic content may pose to women.AFP’s fact-checkers traced hundreds of such videos on Instagram, many in Hindi, that purportedly show male interviewees casually delivering misogynistic punchlines and sexualized remarks — sometimes even grabbing the women — while crowds of men gawk or laugh in the background.Many videos racked up tens of millions of views — and some further monetized that traction by promoting an adult chat app to “make new female friends.”The fabricated clips were so lifelike that some users in the comments questioned whether the featured women were real.A sample of these videos analyzed by the US cybersecurity firm GetReal Security showed they were created using Google’s Veo 3 AI generator, known for hyper-realistic visuals.- ‘Gendered harm’ -“Misogyny that usually stayed hidden in locker room chats and groups is now being dressed up as AI visuals,” Nirali Bhatia, an India-based cyber psychologist, told AFP.”This is part of AI-mediated gendered harm,” she said, adding that the trend was “fueling sexism.”The trend offers a window into an internet landscape now increasingly swamped with AI-generated memes, videos and images that are competing for attention with — and increasingly eclipsing — authentic content.”AI slop and any type of unlabeled AI-generated content slowly chips away at the little trust that remains in visual content,” GetReal Security’s Emmanuelle Saliba told AFP.The most viral misogynistic content often relies on shock value — including Instagram and TikTok clips that Wired magazine said were generated using Veo 3 and portray Black women as big-footed primates. Videos on one popular TikTok account mockingly list what so-called gold-digging “girls gone wild” would do for money.Women are also fodder for distressing AI-driven clickbait, with AFP’s fact-checkers tracking viral videos of a fake marine trainer named “Jessica Radcliffe” being fatally attacked by an orca during a live show at a water park.The fabricated footage rapidly spread across platforms including TikTok, Facebook and X, sparking global outrage from users who believed the woman was real.- ‘Unreal’ -Last year, Alexios Mantzarlis, director of the Security, Trust, and Safety Initiative at Cornell Tech, found 900 Instagram accounts of likely AI-generated “models” — predominantly female and typically scantily clothed.These thirst traps cumulatively amassed 13 million followers and posted more than 200,000 images, typically monetizing their reach by redirecting their audiences to commercial content-sharing platforms.With AI fakery proliferating online, “the numbers now are undoubtedly much larger,” Mantzarlis told AFP.”Expect more nonsense content leveraging body standards that are not just unrealistic but literally unreal,” he added.Financially incentivized slop is becoming increasingly challenging to police as content creators — including students and stay-at-home parents around the world — turn to AI video production as gig work.Many creators on YouTube and TikTok offer paid courses on how to monetize viral AI-generated material on platforms, many of which have reduced their reliance on human fact-checkers and scaled back content moderation.Some platforms have sought to crack down on accounts promoting slop, with YouTube recently saying that creators of “inauthentic” and “mass produced” content would be ineligible for monetization.”AI doesn’t invent misogyny — it just reflects and amplifies what’s already there,” AI consultant Divyendra Jadoun told AFP.”If audiences reward this kind of content with millions of likes, the algorithms and AI creators will keep producing it. The bigger fight isn’t just technological — it’s social and cultural.”burs-ac/st

Foot: accord entre l’OM et l’AC Milan pour Rabiot selon la presse italienne

Marseille et l’AC Milan ont trouvé un accord pour le transfert de l’international français Adrien Rabiot, écarté par l’OM après une altercation avec un coéquipier, rapporte la presse italienne dimanche à la veille de la fin du mercato estival.Selon la Gazzetta dello Sport et la chaîne de télévision Sky Sport, l’AC Milan va débourser dix millions d’euros pour Rabiot qui devrait de son côté signer un contrat de quatre ans avec le club lombard.Contacté par l’AFP, l’AC Milan n’a pour le moment pas donné suite.Une source ayant connaissance du dossier a pour sa part indiqué à l’AFP que le transfert était “presque fait”.A Milan, qui s’est replacé à la huitième place de la Serie A après son succès à Lecce (2-0) vendredi, Rabiot retrouverait Massimiliano Allegri avec qui il a déjà travaillé quand le technicien italien était l’entraîneur de la Juventus Turin. Le recrutement de Rabiot avait été réclamé par Allegri après la défaite à domicile contre la Cremonese (2-1) en ouverture du championnat italien le week-end dernier.L’ancien joueur du Paris SG, 30 ans, connaît bien la Serie A pour y avoir évolué de 2019 à 2024 (157 matches de championnat, 18 buts) sous le maillot de la Juventus, les trois dernières saisons sous la conduite d’Allegri.Rabiot a été mis à l’écart par les dirigeants marseillais après une violente altercation avec un coéquipier, le 15 août à l’issue d’une défaite à Rennes en Ligue 1 (1-0). Il ne s’entraînait plus avec le groupe de Roberto De Zerbi, battu 1 à 0 par Lyon dimanche.Malgré cette situation, Rabiot a été retenu par Didier Deschamps pour les deux premiers matches de qualifications pour le Mondial-2026 contre l’Ukraine et l’Islande, les 5 et 9 septembre.