Timnit Gebru, Mitautorin des Begriffs vom »stochastischen Papagei« für große Sprachmodelle, wird mit dem Right Livelihood Award ausgezeichnet.
Die Informatikerin und prominente KI-Kritikerin Timnit Gebru erhält am Mittwoch den Right Livelihood Award, der auch als Alternativer Nobelpreis bekannt ist.
AI-generated summary
Der Begriff des »stochastischen Papageien« geht auf ein Forschungspapier aus dem Jahr 2021 zurück, an dem Timnit Gebru beteiligt war.
In der Welt der künstlichen Intelligenz (KI) ist es zum geflügelten Wort geworden, die Technik hinter Programmen wie ChatGPT als »stochastischen Papagei« zu bezeichnen. Denn die großen Sprachmodelle lernen aus menschlichen Texten und errechnen dann mithilfe von Statistik plausibel klingende Antworten. Gemeint ist: Die Bots plappern nur nach. Und wir Menschen lassen uns blenden, wenn wir das für Intelligenz halten. Zurück geht der Begriff auf ein Forschungspapier aus dem Jahr 2021. Eine der Autorinnen ist Timnit Gebru. Die Informatikerin ist eine der bekanntesten Kritikerinnen des aktuellen KI‑Hypes. Am Mittwoch erhält sie den Right Livelihood Award, auch bekannt als Alternativer Nobelpreis.

The EU wants to build a Europe-wide, crisis-proof broadband network (EUCCS) for police, fire and rescue services in order to enable cross-border communication and prevent cyber attacks. A gradual introduction is planned until 2036, with Germany pursuing a dual strategy with old and new networks running in parallel. There is criticism of the dependence on private networks and the unclear frequency distribution.

The Munich drone manufacturer Helsing won several simulated dogfights against experienced pilots during a demonstration with an AI-controlled machine. The company is aiming for the first test flight next year.

In parts of its prospectus, Anthropic warns of existential risks to humanity from AI technology. At the same time, AI bosses are trying to convince US President Trump of the dangers, while OpenAI introduces a new AI agent called 'Dots' and holds back another model because of failed security tests.
German air traffic control has put a drone detection system into operation at Berlin Airport BER in order to detect drones in the approach and departure areas at an early stage and to report threats to air traffic. Federal Transport Minister Steffen Bilger described this as an important step for aviation safety. Additional international airports are scheduled to be upgraded over the course of the year.
US President Donald Trump wants to leave the supervision of artificial intelligence to the developers themselves, despite warnings of existential risks. At a meeting with AI company bosses in Washington, he said the companies would monitor each other and provide sufficient protection through their economic interests. A voluntary commitment with principles such as robust internal controls and external observers was agreed upon, signed by executives from Anthropic, OpenAI, Meta, Google and Nvidia.

US President Donald Trump announced that he would leave the supervision of artificial intelligence to the US developers themselves, despite warnings of existential risks. He claimed that a morally binding agreement on self-regulation had already been signed and that there was sufficient protection from the economic interests of companies. Industry representatives such as Dario Amodei from Anthropic, however, warn against uncontrolled AI development and fear a loss of control over the technology. Bipartisan bills to regulate AI exist, but no legislation is likely to be passed before the congressional elections in early November.