Amazon AI images, Google water pledge, Chrome mute update

AI‑generated images on Amazon search

Amazon now injects AI‑created product pictures into its visual‑search results. The images appear when a shopper types a query that matches an item in the retailer’s catalog. Amazon says the move will guide users toward relevant products without requiring them to scroll through text‑only listings.

The feature relies on a visual‑search engine that generates pictures on the fly. TechCrunch reported that Amazon will “show AI generated product images that match your search queries” and that the retailer frames the change as a way to help users find what they need faster. No price or availability data is altered; only the visual cue changes.

Critics note that AI‑generated imagery can mislead shoppers if the rendering diverges from the actual product. The rollout arrives amid broader concerns about deep‑fake content on e‑commerce platforms. Amazon has not disclosed how it will flag synthetic images or give buyers the option to view the original photos.

Google’s water‑use pledge for data centers

Google announced a new water‑stewardship commitment: its data‑center sites will replenish more water than they consume. The pledge expands an existing program and responds to community pushback over the water footprint of AI‑heavy facilities.

9to5Google detailed that the company will track withdrawals and return water through local projects, aiming for a net positive balance. The announcement comes after months of scrutiny about how AI workloads strain municipal supplies, especially in arid regions where cooling towers draw large volumes.

If Google meets the target, it could set a benchmark for other cloud providers. However, the pledge lacks a timeline or independent audit clause, leaving observers to wonder whether the metric will be transparent enough to verify.

Chrome 64 adds permanent mute for autoplay video

Chrome 64 gives users a one‑click way to silence sites that auto‑play videos with sound. The browser adds a permanent mute toggle in the site‑information panel, accessible via the padlock icon at the left of the omnibar.

The update also blocks deceptive links that masquerade as play buttons or close icons. Google says 20 % of desktop feedback reports involve unwanted content, and the new safeguards aim to curb that noise. The mute setting persists until the user manually unmutes the site.

Beyond the mute function, Chrome will soon roll out a Google‑built ad blocker that enforces the Coalition for Better Ads standards. The feature is scheduled for 15 February, indicating a broader push to reduce intrusive experiences without relying on third‑party extensions.

Google tightens content‑quality guidelines

Google revised its Quality Rater Guidelines to label fake news, hate sites, and “monstrously inaccurate” content as low quality. The change, reported by HN, follows a statement from ranking engineer Paul Haahr that only 0.1 % of traffic encounters such results, a figure the company says is disproportionately small compared to media coverage.

The updated guidelines elevate “Your Money or Your Life” pages, demanding the highest E‑A‑T (expertise, authoritativeness, trustworthiness). Raters now have explicit instructions to downgrade sites that spread misinformation or hate, giving Google clearer training data for algorithmic down‑ranking.

Because the affected traffic slice is tiny, Google does not expect immediate changes to its public webmaster guidelines. Nonetheless, the move signals a willingness to codify content standards that could influence future search ranking updates.

Project Zero uncovers lingering Android exploits

Project Zero disclosed an exploit chain for the Pixel 9 that achieved root access from a zero‑click context. The chain leveraged a Dolby vulnerability (CVE‑2025‑54957) that spanned all Android versions until a patch landed in January 2026.

When adapting the exploit for the Pixel 10, researchers found the device replaced the traditional -fstack‑protector with RET PAC, removing the __stack_chk_fail overwrite path. They instead hijacked dap_cpdp_init, an initialization routine called once during decoder setup, to gain code execution.

The findings illustrate how low‑level changes in compiler hardening can shift attack surfaces rather than eliminate them. While Google patched the original flaw, the work underscores the need for continuous scrutiny of emerging mitigations across hardware generations.

What to watch

Track Amazon’s rollout metrics to see whether AI‑generated images boost conversion or spark consumer complaints. Monitor Google’s water‑replenishment reports for independent verification and watch for any regulatory follow‑up. In Chrome, observe user adoption of the mute toggle and the impact of the upcoming ad blocker on ad‑tech revenue. Finally, keep an eye on Google’s next Quality Rater update and any new Project Zero disclosures that could expose residual Android vulnerabilities.