Anthropic Rolls Out a Major Update to Claude's Voice Mode

On July 23, Anthropic announced that it is bringing a more capable model to Claude's Voice Mode. According to the company, the new speech model will make it possible to complete practical tasks — such as rescheduling meetings and drafting emails — entirely by voice, work that previously could only be handled through text.

Competition in voice AI is intensifying, and the focus is on differentiating from existing services such as OpenAI's Advanced Voice Mode and Google's Gemini Live. Anthropic is framing this update around "more natural conversation" and "improved task execution."

ChatGPT's Free Users Get "Inferior Health Advice"

An investigation by The Decoder has revealed that ChatGPT's free users receive lower-quality health advice than paying users. Even for the same health-related questions, the free plan tends to return answers that are more superficial and overly cautious.

This is more than just a feature gap — it exposes an economic disparity in access to medical information. OpenAI only recently rolled out ChatGPT Health nationwide across the US, and attention is now focused on how plan-based differences in answer quality will affect trust in medical AI.

GitHub Cuts Bug Bounty Payouts Amid a Flood of AI-Generated Bug Reports

According to The Register, GitHub has reduced payouts for its public bug bounty after its security team could no longer keep up with a surge of AI-generated bug reports. As automated bug hunting with AI tools has become widespread, the majority of submissions now consist of false positives or low-quality content, leaving vulnerabilities that genuinely need fixing buried in the noise.

The problem reflects the flip side of expectations that AI will improve security. Ensuring the quality of automated reports has become a challenge for bug bounty programs as a whole.

US "AI Kill Switch Act" Draws Attention

The "AI Kill Switch Act," introduced by US House Representative Ted Lieu, is generating debate in the community. The bill aims to mandate the implementation of an emergency shutoff mechanism (a kill switch) for high-risk AI systems.

The details of the bill have been made public as a PDF, but its specific scope and technical requirements will have to wait for further deliberation. As one of the first serious federal-level efforts at AI safety regulation, industry reactions have been mixed.

British Gas Cuts Call Center Staff, Saying "Customers Prefer AI Chatbots"

Chris O'Shea, CEO of Centrica, the parent company of major UK energy supplier British Gas, said that "customers prefer AI chatbots" and revealed that the company is moving forward with cuts to call center staffing. According to the Guardian, the company is accelerating its transition to AI-powered customer service.

The case is drawing attention as a real-world example of AI displacing jobs. At the same time, some criticize the company for offering insufficient evidence behind its customer satisfaction claims and little explanation of how affected employees will be reassigned.

AI Investment Debt Concerns Hit the Tech Bond Market

Corporate bonds issued by major tech companies are being sold off. As Bloomberg reported, concerns over the massive borrowing that underpins the AI boom, combined with a deteriorating situation in the Middle East, are putting downward pressure on the tech bond market.

Meanwhile, an analysis by Finterm takes the view that "Big Tech's $1.65 trillion in AI investment is not debt but an investment in the future," offering a counterargument to market pessimism. The debate over whether AI investment is a "bubble" or "rational forward investment" continues to divide opinion.

---