The world of Artificial Intelligence (AI) is in constant motion. This article offers an overview of current developments, from the increasing digital violence to new AI models and applications in businesses.
The Federal Criminal Police Office (BKA) reports an alarming increase in digital violence against women. Cyberstalking, cybergrooming, and the misuse of AI-generated deepfakes are just a few examples of this concerning development. Organizations like HateAid are demanding tougher action against perpetrators and platforms. The BKA is already implementing measures such as the Central Reporting Office for Criminal Content on the Internet (ZMI BKA) and is cooperating with the Central Office for Combating Internet Crime.
Meta Platforms, the parent company of Facebook, Instagram, and WhatsApp, is establishing a new business unit for B2B AI. Under the leadership of Clara Shih, this unit will develop AI solutions for businesses. With this, Meta is entering a market where Microsoft, OpenAI, and Anthropic are already active. Whether Meta will compete directly with these companies or focus on AI services within the context of its own platforms remains to be seen. Additionally, Meta plans to integrate AI into AR/VR applications.
The Chinese company High-Flyer has introduced DeepSeek-R1, an AI language model with a focus on reasoning. It is intended to compete with models like OpenAI's GPT-4 and achieve comparable performance in benchmarks. However, DeepSeek-R1 is not without flaws and shows weaknesses in complex logical tasks. Noticeable is its refusal to answer political questions about China, presumably to avoid conflicts with the government.
Amazon's plan to upgrade Alexa with AI features is encountering technical difficulties. High latency on older Echo devices and problems integrating third-party services are slowing down development. Internal discussions at Amazon revolve around the question of whether the new AI features should only be made available for newer devices. The concern is that a slower and less reliable Alexa could damage user trust.
Google is equipping its AI chatbot Gemini with a memory function. Users of Gemini Advanced can now store information such as their profession, hobbies, and preferences, which Gemini uses for personalized responses. Similar to ChatGPT Plus, Gemini indicates the use of personal data in its responses. The feature is initially only available in English.
The question of the potential consciousness of AI systems is causing discussion. Experts like Jonathan Birch from the London School of Economics warn of a societal divide due to differing views on AI sentience. Criticism of the tech industry focuses on the lack of interest in the societal consequences. The demand for measurable parameters for AI sentience is growing louder.
Microsoft is launching a bug bounty program with a focus on AI and cloud services. With rewards of up to $30,000, security researchers are encouraged to uncover vulnerabilities in services like Azure, Microsoft 365, and the Power Platform. The program underscores the growing importance of security in the field of AI development.
A survey by the career portal Indeed shows that some respondents prefer working with AI to working with human colleagues. Some even consider AI to be more competent than their colleagues. These results raise questions about the future of collaboration and the impact of AI on the world of work.
Bibliographie: https://www.heise.de/news/KI-Update-kompakt-Digitale-Gewalt-B2B-KI-DeepSeek-R1-Alexa-Gemini-und-mehr-10080291.html https://www.heise.de/newsticker/?from-classic=1 https://www.gender-blog.de/beitrag/ki-und-digitale-gewalt/show https://plus.rtl.de/podcast/ki-update-ein-heise-podcast-fitbmrwnyrukn https://kiupdate.podigee.io/feed/mp3 https://www.youtube.com/watch?v=il39Fif3on4 https://de.linkedin.com/pulse/die-ki-news-der-woche-25-digitale-profis-meixner-ruof-qw0bc https://www.deezer.com/us/show/6138295 https://duepublico2.uni-due.de/servlets/MCRFileNodeServlet/duepublico_derivate_00082171/Ross_ki_und_digitale_gewalt.pdf