Kategori: Genel

  • # OpenAI Bolsters AI Safety with New Biorisk Safeguards for o3 and o4-mini Models

    ## OpenAI Bolsters AI Safety with New Biorisk Safeguards for o3 and o4-mini Models

    OpenAI is taking proactive measures to mitigate potential risks associated with its latest AI reasoning models, o3 and o4-mini. The company announced the deployment of a new monitoring system specifically designed to prevent these models from providing advice that could be used to develop biological or chemical weapons. This safeguard is detailed in OpenAI’s recently released safety report.

    The impetus for this heightened security stems from the enhanced capabilities of o3 and o4-mini compared to their predecessors. OpenAI acknowledges that these models, while not yet crossing the “high risk” threshold, exhibit a greater aptitude for answering questions related to the creation of biological threats. This potential, in the wrong hands, necessitates robust preventative measures.

    The new monitoring system acts as a “safety-focused reasoning monitor,” operating on top of the o3 and o4-mini models. It is custom-trained to understand and enforce OpenAI’s content policies, specifically targeting prompts related to biological and chemical hazards. When such prompts are detected, the monitor instructs the AI to refuse to provide assistance.

    To establish a strong foundation for this system, OpenAI’s “red teamers” dedicated approximately 1,000 hours to identifying and flagging unsafe, biorisk-related conversations generated by the models. Subsequent testing, simulating the monitor’s blocking logic, demonstrated a 98.7% success rate in preventing the models from responding to risky prompts.

    However, OpenAI recognizes the limitations of automated systems. The company acknowledges that determined individuals might attempt to circumvent the monitor by crafting new, unforeseen prompts. To address this, OpenAI will continue to rely on human monitoring alongside the automated safeguards.

    This initiative underscores OpenAI’s growing commitment to AI safety, as evidenced by its recently updated Preparedness Framework. The company is actively tracking potential misuse scenarios and implementing automated systems to mitigate risks. A similar reasoning monitor is also being used to prevent GPT-4o’s native image generator from creating harmful content, such as child sexual abuse material (CSAM).

    Despite these efforts, some researchers remain concerned that OpenAI isn’t prioritizing safety adequately. Critics point to limited testing time for certain benchmarks and the absence of a safety report for the recently launched GPT-4.1 model.

    While the debate around AI safety continues, OpenAI’s proactive steps to address potential biorisks associated with its latest AI models demonstrate a growing awareness of the need for robust safeguards in an increasingly powerful and rapidly evolving AI landscape. As AI models continue to advance, the ongoing development and refinement of these safety mechanisms will be crucial to ensure responsible and ethical deployment.

  • # Rippling, Deel’in İddia Edilen “Casusunu” Kimin Ödediğini Açıklaması İçin Revolut’a Baskı Yapıyor

    ## Rippling, Deel’in İddia Edilen “Casusunu” Kimin Ödediğini Açıklaması İçin Revolut’a Baskı Yapıyor

    Rippling ve Deel arasındaki devam eden hukuk dramasına, İngiliz fintech devi Revolut da dahil oldu. Bu karmaşık dava, teknoloji dünyasında büyük yankı uyandırmaya devam ediyor.

    Olaylar, Rippling çalışanı Keith O’Brien’ın, Deel tarafından Rippling’i gözetlemek için ayda yaklaşık 6.000 dolar ödendiğini iddia ettiği bir ifadeyle başladı. O’Brien’ın iddiasına göre, ilk 6.000 dolarlık ödeme Kasım 2024’te Deel’in COO’su Dan Westgarth’ın eşi Alba Basha Westgarth tarafından Revolut hesabına gönderildi.

    TechCrunch tarafından incelenen ve daha sonra silinen bir LinkedIn profiline göre Alba Basha Westgarth, o dönemde Robinhood’da kripto uyum lideri olarak görev yapıyordu. Robinhood, TechCrunch’a Westgarth’ın bu yılın başlarında şirketten ayrıldığını doğruladı, ancak ayrılış nedenine ilişkin bir açıklama yapmadı.

    Silinen LinkedIn profilinde Westgarth’ın konumunun Dubai olarak belirtilmesi dikkat çekici. Zira Rippling’in CEO’su ve hukuk direktörü de şu anda Dubai’de bulunuyor ve Rippling’in onlara tebligat yapma çabaları sonuçsuz kalıyor. Birleşik Arap Emirlikleri’nin iade konusunda çekimser bir ülke olarak bilinmesi, olaya ayrı bir boyut katıyor.

    Rippling, şimdi de o 6.000 dolarlık makbuzda gönderici olarak listelenen “Alba Basha”nın tam adını ve adresini öğrenmek için Revolut’a dava açtı. Mahkeme kayıtlarına göre Rippling, Alba Basha’nın kimliğini kesin olarak doğrulamak için kimlik ve fatura gibi tüm hesap açılış belgelerini de talep ediyor.

    Revolut ise bu talebe yanıt olarak İrlanda’nın önde gelen bir avukatını görevlendirdi. Ayrıca, Rippling’in avukatının mahkeme duruşmasında “yardımcı ama karmaşık” olarak tanımladığı bir mektubu da Rippling’e gönderdi. Ancak avukat, mektubun içeriği hakkında daha fazla detay vermedi.

    Burada açıkça belirtmek gerekir ki, Revolut’un herhangi bir yanlış davranışta bulunduğuna dair bir işaret veya suçlama bulunmuyor. Avrupa Birliği’nin katı gizlilik yasaları nedeniyle, bu tür bilgileri sağlamak özellikle bu gibi hukuk davalarında kolay değil. Temel sorunlardan biri, Revolut’a bilgi sağlama konusunda İrlanda mahkemelerinden herhangi bir emir gelmemiş olması.

    Revolut, dava hakkında özel bir yorum yapmaktan kaçınırken, genel olarak “açıklama talebi içeren her türlü mahkeme emrine uyacağını” belirtti.

    Transferin arkasındaki “Alba Basha” hakkında Revolut’un daha fazla ayrıntı sağlayıp sağlamayacağı belirsizliğini koruyor. Avrupa gizlilik yasaları kapsamındaki tek diğer seçenek, müşterinin onayı olacaktır ki bu da bu durumda pek olası görünmüyor.

    Deel ise medyaya yaptığı açıklamalarda herhangi bir yanlış davranışta bulunmadığını iddia ediyor. Deel COO’su Dan Westgarth ve Alba Basha Westgarth ise yorum taleplerine yanıt vermedi. Rippling de konu hakkında yorum yapmaktan kaçındı.

    Kanıtlar henüz kesin olmasa da, İrlanda Yüksek Mahkemesi, Alba Basha’nın kimliğini bir şekilde doğrulama konusunda Rippling’in dileklerini kabul edebileceğine dair bir ipucu verdi.

    Rippling’in avukatının, 2 Nisan’daki bir duruşmada Alba Basha’yı Dan Westgarth’a bağlayan kanıtları (ikilinin birlikte olduğu bir Facebook fotoğrafı da dahil olmak üzere) sıralamasının ardından, Yargıç Mark Sanfey, “Tesadüf olabilir, ama pek olası değil!” yorumunu yaptı.

    Bu olay, teknoloji dünyasındaki rekabetin ne kadar acımasız olabileceğini ve şirketlerin rekabet avantajı elde etmek için ne kadar ileri gidebileceğini gösteriyor. Dava, aynı zamanda kişisel verilerin korunması ve uluslararası hukuk gibi önemli konuları da gündeme getiriyor. Gelişmeleri yakından takip etmeye devam edeceğiz.

  • # Krebs to Challenge Trump-Ordered Investigation, Cites Punitive Dissent Tactics

    ## Krebs to Challenge Trump-Ordered Investigation, Cites Punitive Dissent Tactics

    Former Cybersecurity and Infrastructure Security Agency (CISA) Director Chris Krebs is set to fiercely contest a federal investigation launched against him, initiated by former President Trump. In an interview with *The Wall Street Journal* on Wednesday, Krebs announced his intention to resign from his current position at cybersecurity firm SentinelOne to dedicate his efforts to challenging the probe.

    The investigation centers on accusations that Krebs “falsely and baselessly denied that the 2020 election was rigged and stolen.” The order also revoked Krebs’ security clearance. Krebs contends that the investigation is a politically motivated attack designed to suppress dissent and target corporate relationships. He sees the action as an attempt by the government to wield its power to punish those who disagree with its narratives.

    Krebs, appointed by Trump in 2018 to head the newly established CISA, was responsible for overseeing federal cybersecurity and election security. However, his tenure was cut short in November 2020 after he publicly refuted Trump’s unfounded claims of widespread voting irregularities. This dismissal followed CISA’s efforts to debunk misinformation surrounding the election, earning Krebs widespread praise from cybersecurity experts and condemnation from Trump allies.

    According to *The Wall Street Journal*, Krebs is joining a growing coalition of former government officials, law firms, and universities who are pushing back against what they see as the Trump administration’s abuse of executive power to silence critics and stifle dissenting voices. The case highlights the ongoing debate regarding the separation of powers and the potential for politically motivated investigations to undermine public trust and free speech.

  • # US Weighs DeepSeek Ban Amid AI Competition with China

    ## US Weighs DeepSeek Ban Amid AI Competition with China

    The Trump administration is reportedly considering a ban on Chinese AI lab DeepSeek, a move that could significantly impact the AI landscape in both the US and China. According to a report in *The New York Times*, the proposed restrictions could prevent DeepSeek from purchasing Nvidia’s advanced AI chips and potentially block Americans from accessing the company’s AI services.

    This potential ban is part of a broader strategy by the Trump administration to bolster the US’s position in the global AI race. DeepSeek has rapidly gained prominence, becoming a key competitor in the AI model market. Its competitive pricing and advanced capabilities have spurred Silicon Valley companies to lower the costs of their own frontier AI models. This disruption has prompted US officials to explore various methods of limiting China’s access to American technology and consumers.

    The administration’s focus on restricting access to AI chips isn’t new. Just this week, the White House expanded existing restrictions on Nvidia’s AI chip sales to China, building upon rules initially established by the Biden administration. These measures aim to slow down China’s progress in AI development by limiting its access to crucial hardware.

    However, the potential ban also raises questions about DeepSeek’s competitive advantage. OpenAI has previously accused the Chinese lab of potentially violating its terms of use by distilling its models. These allegations of intellectual property theft add another layer of complexity to the situation, prompting further scrutiny of DeepSeek’s practices. The outcome of this consideration could reshape the dynamics of the AI industry, influencing innovation and competition in the years to come.

  • # Trump Yönetimi Çinli AI Laboratuvarı DeepSeek’e Yasak Getirmeyi Düşünüyor

    ## Trump Yönetimi Çinli AI Laboratuvarı DeepSeek’e Yasak Getirmeyi Düşünüyor

    Trump yönetimi, Çinli yapay zeka laboratuvarı DeepSeek’e yönelik yeni kısıtlamalar getirmeyi değerlendiriyor. New York Times’ın haberine göre bu kısıtlamalar, DeepSeek’in Nvidia’nın yapay zeka çiplerini satın almasını engelleyebilir ve Amerikalıların şirketin yapay zeka hizmetlerine erişimini kısıtlayabilir.

    Bu hamle, Trump yönetiminin Çin ile yapay zeka alanındaki rekabetini artırma çabalarının bir parçası olarak değerlendiriliyor. DeepSeek’in Silikon Vadisi ve Wall Street’i sarsmasının ardından, ABD’li yetkililer Çin’in Amerikan teknolojilerine ve tüketicilerine erişimini sınırlamak için çeşitli seçenekleri değerlendiriyor gibi görünüyor.

    Beyaz Saray, Salı günü Nvidia’nın Çin’e yapay zeka çip satışlarına yönelik kısıtlamaları daha da sıkılaştırdı. Bu hamle, Biden yönetimi tarafından oluşturulan kuralları güçlendiriyor.

    Son aylarda, DeepSeek’in ABD’li yapay zeka geliştiricileri arasındaki popülaritesi hızla arttı. Startup’ın rekabetçi fiyatlandırması, Silikon Vadisi’ni daha uygun maliyetli ileri düzey yapay zeka modelleri sunmaya zorladı.

    Ancak DeepSeek’in bazı rekabetçi modellerini oluştururken fikri mülkiyet hırsızlığı yapıp yapmadığı konusunda süregelen sorular var. OpenAI, Çinli laboratuvarın kendi modellerini “damıttığını” ve OpenAI’nin kullanım şartlarını ihlal ettiğini iddia ediyor.

  • # Rippling Intensifies Legal Battle with Deel, Subpoenas Revolut for “Spy” Payment Information

    ## Rippling Intensifies Legal Battle with Deel, Subpoenas Revolut for “Spy” Payment Information

    The ongoing legal clash between HR tech companies Rippling and Deel has taken a dramatic turn, drawing in UK fintech giant Revolut. Rippling is seeking to uncover the identity of an individual who allegedly paid a Rippling employee to spy on the company for Deel.

    The crux of the issue lies in an affidavit from Keith O’Brien, an Irish Rippling employee, who claims he was paid approximately $6,000 per month by Deel to gather intelligence. O’Brien stated that the initial payment was routed through his Revolut account, with the sender listed as “Alba Basha.” According to O’Brien, this Alba Basha is Alba Basha Westgarth, the wife of Deel’s COO, Dan Westgarth.

    Adding another layer of intrigue, Alba Basha Westgarth previously held the position of crypto compliance lead at Robinhood, according to a LinkedIn profile that has since been deleted. Robinhood confirmed her departure earlier this year but declined to provide a reason. Furthermore, the deleted profile listed her location as Dubai, the same location where Deel’s CEO and legal director are currently based, complicating Rippling’s efforts to serve them papers. Dubai is notably known for its lack of extradition treaties.

    Rippling has now filed a lawsuit against Revolut, aiming to compel the fintech company to disclose the full name and address associated with the “Alba Basha” account that sent the $6,000 payment. They are also requesting all account opening documentation, including identification and utility bills, to definitively confirm her identity.

    Revolut has responded by hiring legal counsel in Ireland and providing Rippling with a letter described by Rippling’s legal team as “helpful but complicated.” While there are no accusations of wrongdoing against Revolut, the case highlights the complexities of complying with data requests under stringent EU privacy laws. Disclosing customer information requires either a court order from the Irish court or the customer’s explicit consent.

    Revolut issued a statement confirming its commitment to complying with legitimate court orders for disclosure but did not comment directly on the Rippling case. As the legal battle unfolds, the question remains whether Revolut will be compelled to provide the requested information. Customer consent appears unlikely in this scenario.

    Deel, which has consistently denied any involvement in espionage, has not responded to requests for comment. Neither Deel COO Dan Westgarth nor Alba Basha Westgarth have issued statements. Rippling has also declined to comment.

    Despite the lack of definitive proof, a judge in the Irish High Court hinted that Rippling’s request to verify Alba Basha’s identity might be honored. This came after Rippling’s lawyer presented evidence linking Alba Basha to Dan Westgarth, including a Facebook photo of the pair. The judge acknowledged that the connection “could be a coincidence, but it’s unlikely!” The Rippling-Deel legal saga continues to develop, with Revolut now caught in the crossfire.