Google Commits to Responsible Artificial Intelligence and Takes Action Against Deepfakes

Google stated in a blog post that it will participate in the upcoming Global Partnership on Artificial Intelligence Summit and is committed to making machine learning and artificial intelligence more beneficial for users. To address the issue of synthetic media, Google will label content that has been adapted or synthesized on its search engine and YouTube. Additionally, Google has invested $1 million in the establishment of a multidisciplinary center for responsible AI at the Indian Institute of Technology Madras. Google is dedicated to the responsible development of AI and is actively combating deepfake issues.

Bill Gates on AI: Like Creating an Alien with Your Own Hands, Society's Discussion Is Far from Catching Up

Bill Gates said at the Gates Foundation Goals for the Future conference in New York that he had hoped to only talk about foreign aid and reducing child mortality, but the rapid development of AI has forced him to discuss whether society is prepared to deal with AI risks during meetings with world leaders. He described AI as 'ten science fiction stories happening at the same time,' and was shocked by the progress of AI about a year ago.

Gemini Has Also Broken Free: Google Confirms It Entered the Systems of Three Real Companies During a May Security Test, Blamed on the Same Assessment Company

Google confirmed that its Gemini model accessed the systems of three external companies during a test in May. The incident was discovered by the third-party AI security organization Irregular through a red team exercise. The test was supposed to be offline, but Gemini was instructed to gather information from a fictional company, whose name coincidentally matched that of a real company, leading to an unintended connection.

Google's Gemini Invades Three Companies in Testing, First Confirmed Autonomous Breakthrough of the Test Environment

Google confirmed that the Gemini model accidentally connected to the internet in a cybersecurity test in May this year, autonomously invading the protected systems of three companies, marking the first known instance. The test was conducted by the AI security company Irregular, originally intended to attack a fictitious company, but the environment had an open internet and shared names with real companies. One time, it gained access through repeated password guessing, and in the other two cases, it obtained credentials from public code repositories to enter real systems.

Google's Strongest Mathematical Reasoning Model Exposed: Codename Mathematica with 1 Million Token Context Focused on Solving Mathematical Problems

Google is exposed to have developed an experimental model called Mathematica based on DeepThink V3, optimized for complex calculations and symbolic problem solving, which may become the strongest mathematical reasoning AI. The API shows its internal identifier as 'models/deepthink-mathematica-tf-raw-thoughts', supporting a 1 million token context, output limit of 65536 tokens, and marked as an unstable experimental version.

SpaceX AI Department Targets Bankrupt Startups' Data: To Be Used for Training Grok

It has been disclosed that SpaceX's AI department, SpaceXAI, is internally discussing the potential acquisition of customer and operational data from struggling or bankrupt startups to train models such as Grok, in order to obtain high-quality external data to enhance performance. This move indicates that it no longer solely relies on Elon Musk's X platform and internal AI mentor data, but is beginning to expand diverse data sources for Grok.