av Mikael Winterkvist | dec 26, 2025 | Bluesky, Mastodon, Notiser, Threads
OpenAI acknowledges that prompt injections – text-based attacks on language models running in browsers – may never be completely eliminated. Still, the company says it’s ”optimistic” about reducing the risks over time.
OpenAI has released a security update for the browser agent in ChatGPT Atlas. The update includes a newly adversarially trained model and enhanced security measures, prompted by a new class of prompt injection attacks discovered through OpenAI’s internal automated red-teaming.
Källa: OpenAI admits prompt injection may never be fully solved, casting doubt on the agentic AI vision
av Mikael Winterkvist | dec 26, 2025 | Bluesky, Mastodon, Notiser, Threads
The spread of always-on cameras has created a privacy problem where bystanders lose control over how and when they are captured on video.
To address this, researchers at the University of California, Irvine designed BLINDSPOT, an on-device privacy signaling system. It allows bystanders to communicate privacy preferences to camera-enabled devices without identity registration, biometric uploads, or cloud processing. The prototype was implemented and evaluated on a Google Pixel smartphone.
BLINDSPOT gives people in a camera’s field of view an explicit way to control how their faces appear in recorded video. When a bystander sends a signal, the device detects and tracks that person’s face and applies blurring before the video is stored or shared.
Källa: What if your face could say ”don’t record me”? Researchers think it’s possible – Help Net Security
av Mikael Winterkvist | dec 25, 2025 | Bluesky, Mastodon, Notiser, Threads
Nearly two years ago, Donald Trump kicked off the presidential-campaign season with a declaration: “I was never on Epstein’s Plane, or at his ‘stupid’ Island,” he posted on Truth Social in January 2024. Reports to the contrary, he insisted, were the fault of AI—and of his political rivals: “This is what the Democrats do to their Republican Opponent, who is leading them, by a lot, in the Polls.”
But this week, the documents released by Trump’s own Justice Department—including flight logs and emails—told a different story. Federal prosecutors determined in January 2020 that Trump had been a passenger on the notorious private jet owned by Jeffrey Epstein—who would later be charged with sex trafficking—far more often than they had realized.
Källa: So This Is Why Trump Didn’t Want to Release the Epstein Files
av Mikael Winterkvist | dec 25, 2025 | Bluesky, Mastodon, Notiser, Threads
Cloudflare has released the 2025 Cloudflare Radar Year in Review, highlighting the trends it observed throughout the year as one of the world’s largest internet infrastructure providers. Here’s how Apple did.
Apple-specific trends in 2025, per Cloudflare’s report
According to Cloudflare (via MacMagazine), Apple ranked third in the Internet Services ranking, which ranks major online services by relative popularity using Cloudflare’s global DNS data.
Cloudflare’s data shows that Apple overtook TikTok early in the year, and stood firmly in third place until around July, when it began switching places with Microsoft.
Källa: Apple lands third in Cloudflare’s 2025 Internet Services rankings – 9to5Mac
av Mikael Winterkvist | dec 25, 2025 | Bluesky, Mastodon, Notiser, Threads
The release of a new tranche of files related to Jeffrey Epstein will only add to Americans’ confusion and division over the late sex offender and his ties to Donald Trump and other elite figures, analysts said.
The thousands of documents US authorities released on Tuesday contain many mentions of Trump, but also include random, unverified tips that investigators had received, as well as files that the Department of Justice said were fabricated.
Some of the pages, meanwhile, are heavily redacted for undisclosed reasons. The release has “fanned the flames of the conspiracy,” while also “muddying the waters” with unverified claims, The Atlantic’s Charlie Warzel wrote.
Källa: US authorities release new tranche of Epstein files
av Mikael Winterkvist | dec 25, 2025 | Artificiell intelligens, Bluesky, Mastodon, Threads

Efter att länge ha hållit dörren stängd för annonser ser OpenAI nu ut att byta fot. Enligt nya uppgifter från The Information arbetar företaget för fullt med att utveckla olika reklamformat för ChatGPT. Det handlar om allt från diskreta sidofält till mer kontroversiella metoder där sponsrat innehåll vävs direkt in i AI-modellens svar.
Sponsrade svar och smarta länkar De interna skisserna visar flera tänkbara scenarier. Om en användare ber om rekommendationer på mascara, skulle svaret kunna prioriteras så att ett specifikt märke lyfts fram. En annan modell innebär att reklam dyker upp först när användaren ber om mer detaljer. Den som planerar en resa till Barcelona och klickar på ett förslag om Sagrada Família kan då få upp sponsrade länkar till guidade turer och resepaket.
Från dystopi till affärsmodell Det mest anmärkningsvärda är att OpenAI verkar vilja använda chattens minnesfunktion för att rikta annonserna. Detta innebär att tidigare, privata konversationer kan ligga till grund för vilken reklam som visas. Det är ett steg som går stick i stäv med vad vd Sam Altman tidigare har sagt. Han har nämligen kallat en framtid där AI-svar formas av annonser för ”dystopisk”, särskilt om de bygger på användarens personliga historik.
En talesperson för OpenAI bekräftar för The Information att de utforskar hur reklam skulle kunna fungera i produkten, men understryker att de vill göra det utan att skada användarnas förtroende. Frågan är dock hur väl mottaget det blir när användarna inser att deras privata chattar kan förvandlas till underlag för riktad marknadsföring.
Källa: The Decoder