Starting in October, Microsoft plans to bring back its AI-powered memory tool, Recall, for participants in the Windows Insider Program who are using Copilot+ PCs. Recall captures screenshots regularly, analyzing and correlating them using AI, and making the results searchable. This feature raised privacy concerns, prompting Microsoft to initially offer it as an opt-in feature while working on encryption and privacy enhancements.

Originally, Recall was supposed to be enabled by default on all Copilot+ PCs. However, after facing criticism, it was turned into an opt-in feature that users must activate themselves. Additionally, screenshots will now be stored in an encrypted format, and searches will only be possible after confirmation via Windows Hello. In managed environments, administrators will only be able to disable the feature centrally, not enable it. Microsoft emphasized that security remains the highest priority and that more details will be revealed when Recall becomes available to Windows Insiders in October.

Google DeepMind Employees Demand End to Military Contracts

At Google DeepMind, the AI-focused subsidiary of Google in the UK, discontent is growing over the company’s involvement in supporting big data solutions for military purposes. Around 200 employees have signed a letter urging Google to terminate its contracts with military organizations. The issue centers on reports that Google has a contract to provide cloud computing and AI services to the Israeli military and is part of Project Nimbus, which has been linked to mass surveillance and target selection during the Gaza conflict.

The signatories argue that any involvement in the military and weapons production undermines DeepMind’s position as a leading company in ethical and responsible AI, violating its mission and stated AI principles. They are calling for military users to be denied access to DeepMind technologies and for an oversight committee to be established to ensure this.

Google has yet to respond to these demands, leading to growing frustration among employees. When Google acquired DeepMind in 2014, it pledged never to use AI technology for military or surveillance purposes. While Google insists it adheres to its AI principles, it has not ruled out serving military clients with its AI.

Google Avoids Law by Agreeing to Pay Millions to Publishers in California

To avoid a proposed law in California that would have required Google to pay publishers for using their content and links, the company has committed to voluntary payments totaling hundreds of millions of dollars. Over the next five years, Google will provide $250 million to local newsrooms. The deal also includes the development of an AI program designed to assist journalists in their work. The state of California will also contribute to journalism and related initiatives.

Journalists’ representatives, who were excluded from the negotiations, have criticized the agreement. They had supported the original legislative proposal and describe the deal as a “catastrophic agreement” that supports monopolies. They argue that the future of journalism should not be negotiated behind closed doors. Earlier this year, Google stopped linking to California news outlets in protest of the proposed law, calling it a test to assess the law’s impact.

Apple’s Plans for a Robotic Arm with AI Personality

According to Bloomberg reporter Mark Gurman, Apple is currently exploring the field of robotics as part of its strategy to expand its presence in consumers’ homes and introduce a new dimension to its product lineup. However, the company is still in the early stages and needs to determine the best applications for robotics. A device codenamed J595 is serving as a test case, which is expected to be released around 2026 or 2027. It combines a large, iPad-like display with cameras and a base featuring a robotic actuator. In the next decade, Apple could introduce mobile robots and possibly even humanoid models.