Microsoft and Google Defend AI Quality

Within the space of a single week, senior figures from Microsoft and Google shared views on growing criticism around artificial intelligence, offering similar perspectives on how the debate should be framed. Rather than focusing on whether AI outputs are high or low quality, both shifted attention towards how users are responding to the technology itself.

Microsoft chief executive Satya Nadella and Jaana Dogan, a Principal Engineer working on Google’s Gemini API, each commented publicly on the issue. Their remarks arrived at a time when publishers continue to report falling website traffic linked to AI-driven search features.

Nadella shared his thoughts in a personal blog post titled Looking Ahead to 2026. In it, he argued that the industry should move beyond the ongoing debate that labels AI content as either “slop” or “sophisticated”.

He suggested that this framing oversimplifies the conversation and distracts from more important questions about how artificial intelligence fits into everyday work and life. Nadella described AI tools as “cognitive amplifiers” and said the coming years will be critical in proving their real-world value.

According to Nadella, the focus should not be on ending discussion altogether, but on reshaping it. He called for what he described as a “new equilibrium”, one that recognises how humans and AI tools operate together. In his view, this is fundamentally a product design challenge that still needs to be debated and resolved.

Only days later, Jaana Dogan shared her own thoughts on social media, offering a different but related interpretation of resistance to AI. She suggested that negative attitudes towards new technology often come from exhaustion rather than outright opposition.

In her post, Dogan wrote that people tend to become “anti new tech” when they are worn down from repeatedly testing and adapting to emerging tools. She described this reaction as understandable, particularly in fast-moving technical environments.

Her comments followed an earlier post in which she described using Anthropic’s Claude Code to build a functional prototype based on a short technical description. Dogan said the tool produced results in around an hour that closely resembled systems her team had spent nearly a year developing.

She added that just two years ago, she believed these capabilities were still several years away, highlighting how quickly AI tools have advanced in a short period of time.

However, Dogan’s remarks about burnout sparked criticism from other users online. Many responses argued that frustration with AI is not simply fatigue, but stems from forced rollouts, rising costs, privacy concerns, and tools that do not always integrate smoothly into daily workflows.

It is also worth noting that Dogan’s comments were shared in a personal capacity. While she works on Google’s Gemini API, she was not speaking as an official representative of Google’s corporate position.

For publishers, these statements land against a backdrop of long-standing quality expectations imposed by large platforms. For years, Google has emphasised the importance of experience, expertise, authority, and trustworthiness, particularly for sensitive topics such as health, finance, and legal advice.

This is why the current situation feels contradictory to many content creators. While publishers are expected to meet strict quality standards, AI products increasingly deliver information directly to users, often with citations that are difficult to assess quickly.

When questioned about falling click-through rates, platform representatives have spoken publicly about “higher-quality clicks”, rather than addressing the overall decline in traffic volumes reported by publishers.

Data from the Pew Research Center highlights the scale of the issue. After analysing nearly 69,000 Google searches, researchers found that when AI-generated summaries appeared, only 8% of users clicked on a link. Without AI summaries, that figure rose to 15%, representing a drop of almost 47%.

While remaining clicks may show stronger intent, volume still matters. Traffic underpins advertising revenue, subscriptions, and affiliate income, all of which are critical to the sustainability of online publishing.

Further figures from Similarweb show that the share of news-related searches ending without any click-through to a news website increased from 56% to 69%, reinforcing concerns about reduced visibility.

There is also growing attention on the imbalance between crawling and referrals. Cloudflare estimates suggest that Google Search operates at roughly a 14-to-1 crawl-to-referral ratio, while AI-focused platforms show dramatically higher figures, with OpenAI and Anthropic far exceeding that level.

For many publishers, the long-standing trade-off has been simple: allow platforms to crawl content in exchange for visibility and traffic. Increasingly, there is concern that AI features weaken this arrangement by answering queries directly without sending users back to original sources.

Taken together, the comments from Nadella and Dogan offer a glimpse into how debates around AI quality may be handled in 2026. By framing criticism as a reaction problem rather than a product or economic issue, the discussion risks moving away from accuracy, reliability, and financial impact.

Traffic declines are measurable, referral imbalances are visible, and the consequences for publishers are tangible.

Looking ahead, more messaging may emerge that positions AI criticism as user resistance rather than a reflection of unresolved design and economic challenges. Whether this leads to meaningful changes in how AI products are built and deployed remains to be seen.