TL;DR: Google’s search quality began to feel “weird” to many users around 2022-2024 as AI integration accelerated, algorithm updates prioritized ads and AI-generated content, and features like AI Overviews produced bizarre results. This article traces the key events that shifted Google from a reliable search tool to an experimental AI playground, leaving Malaysians and global users questioning the reliability of their go-to search engine.
The Ads Influence: When Priorities Shifted
Google’s search engine has long been funded by advertising, but around 2020, the company placed its head of Ads in charge of Search, marking a significant shift in priorities. As reported by industry analysts, this move allowed Ads department objectives to creep into search algorithms, leading to complaints that “Google search is getting worse” becoming a common refrain. The fundamental tension emerged: should Search serve users’ information needs or advertiser interests? Users began noticing more branded content and less diverse, independent sources in their results, setting the stage for future frustrations.
Specific changes included an increase in above-the-fold advertisements, pushing organic results further down the page. Shopping ads became more prominent for product queries, often occupying the entire first screen on mobile devices. Local service ads appeared for queries like “plumber near me” or “electrician Kuala Lumpur,” sometimes before the traditional local pack. These shifts meant users had to scroll more to find organic results, increasing frustration especially on mobile devices where screen real estate is limited.
Furthermore, the Ads influence affected ranking signals indirectly. As Google’s internal teams prioritized revenue-generating features, resources for quality-focused ranking improvements were sometimes diverted. This created a perception that the search engine was becoming more commercialized and less user-centric, laying groundwork for later disappointments when AI integration amplified these tensions.
The Helpful Content Update: A Well-Intentioned Shift
In August 2022, Google launched the Helpful Content Update (HCU) with the goal of “better connecting people to helpful information.” The update encouraged publishers to create “people-first content” and avoid making content primarily for search engines. The original guidance emphasized content “written by people, for people,” aiming to reward original, helpful content while demoting low-effort, AI-generated material. For a brief period, this update was welcomed as a potential guardrail against the impending wave of AI-generated content following ChatGPT’s late 2022 debut.
The HCU arrived at a pivotal moment. Website owners and SEO professionals had been grappling with an influx of low-quality, keyword-stuffed content designed to game rankings. Google’s message was clear: focus on creating genuine value for human readers rather than chasing algorithmic tricks. Many publishers embraced this shift, investing in deeper research, expert authorship, and user experience improvements. Early data showed promising reductions in certain types of spammy content, giving hope that Google could maintain search quality despite rising AI capabilities.
However, the HCU also revealed limitations in Google’s ability to assess content quality algorithmically. The update relied heavily on user engagement signals and site-wide assessments, which sometimes penalized legitimate sites with niche audiences or unconventional formats. Critics noted that big brands often benefited disproportionately due to stronger brand signals, while smaller independent creators struggled to gain traction despite producing high-quality content. These early shortcomings foreshadowed the challenges that would emerge when AI entered the equation.
The AI Race Begins: Bard, Gemini, and Conflicting Goals
Just as the HCU was taking effect, Google entered the AI race with Bard in early 2023, later rebranded as Gemini. This move placed Google directly in the content creation business, creating a conflict of interest: the same company ranking content was now producing it. As one tech analyst noted, “Just as when Ads’ priorities crept into the search algorithms, Google’s search results are threatened by priorities driven by its AI business.” For the first time in its history, Google found itself playing catch-up in search while trying to dominate AI, leading to inconsistent messaging about what constituted helpful content.
The timing was particularly delicate. ChatGPT had captured public imagination in late 2022, sparking concerns about a flood of AI-generated content that could overwhelm search results. Google’s initial HCU launch had positioned it as a defender of human-written content. Yet within months, the company was promoting its own AI tools as beneficial for content creation. This contradiction created confusion among publishers: should they avoid AI content as potentially harmful to rankings, or embrace Google’s AI tools to stay competitive? The lack of clear guidance exacerbated uncertainty during a critical transition period.
Policy Reversals: From “Written by People” to “Created for People”
A year after the HCU, Google quietly updated its guidance, dropping the phrase “written by people” in favor of “created for people.” This subtle change signaled a shift in stance toward AI-generated content. Critics pointed out that the original wording clearly stood against machine-spun content, and its removal suggested Google was softening its position to accommodate its own AI ambitions. The updated policy stated that automation wasn’t inherently problematic—what mattered was quality—effectively giving a green light to scaled AI content production as long as it was deemed helpful.
This policy shift had immediate consequences. Content farms and SEO agencies began mass-producing AI-generated articles targeting long-tail keywords, confident that Google’s updated stance meant such content would rank well if it met basic quality thresholds. The result was a surge in low-effort, AI-assisted content that often lacked depth, originality, or factual accuracy. While some pieces were genuinely helpful, many existed primarily to capture search traffic rather than provide meaningful value. Users began encountering repetitive, generic answers across multiple sites, diminishing the diversity of perspectives that had made Google’s search results valuable.
The HCU Folded Into Core Algorithm: March 2024
In March 2024, Google folded the Helpful Content Update system into its core ranking algorithm. This technical change came with a policy update: AI-generated content was no longer automatically considered a red flag. Instead, Google’s updated spam policies targeted content produced at scale to manipulate rankings, regardless of whether it was made by humans, machines, or both. While this approach seemed balanced, it inadvertently encouraged mass production of AI content that often lacked originality or insight, leading to a flood of low-value pages that frustrated users seeking genuine information.
The folding of the HCU into the core algorithm represented a philosophical shift. Rather than treating helpful content as a separate, special signal, Google integrated its principles into the broader ranking system. This meant that helpfulness became just one factor among many, potentially diluted by other considerations like page speed, mobile-friendliness, and backlink profiles. For publishers, the change meant that HCU compliance alone was no longer sufficient for good rankings; they needed to excel across all ranking factors simultaneously.
More significantly, the updated spam policy’s focus on “content produced at scale to manipulate rankings” created a loophole. AI-generated content that demonstrated some level of helpfulness or user engagement could escape punishment, even if it was clearly created primarily for search traffic. This ambiguity led to a proliferation of “helpful-seeming” AI content that satisfied superficial quality metrics while failing to deliver substantive value. Users searching for complex topics often found pages that appeared informative at first glance but contained shallow, regurgitated information upon closer inspection.
AI Overviews Launch: Glue on Pizza and Eroded Trust
May 2024 saw the launch of AI Overviews (AIOs), Google’s Gemini-powered search summaries designed to answer queries directly in the results page. The rollout was immediately problematic, with users sharing screenshots of bizarre and incorrect answers: suggestions to put glue on pizza to make cheese stick, advice to eat rocks for nutrients, and other hallucinations that quickly went viral. These errors undermined public trust in Google’s search capabilities, especially as AI Overviews began appearing prominently for many queries, pushing down traditional organic results that users had relied on for years.
Specific incidents highlighted the risks of premature AI deployment. A query about “cheese not sticking to pizza” returned an answer suggesting non-toxic glue as a solution—a dangerous recommendation that could have caused harm if followed. Medical queries produced equally alarming results, with AI Overviews suggesting harmful home remedies for serious conditions. Even seemingly harmless queries about historical facts or celebrity biographies sometimes returned confidently stated falsehoods. These errors weren’t just embarrassing; they posed real risks to users who might trust AI-generated answers without verification.
The impact extended beyond individual user experiences. Publishers whose content was summarized by AI Overviews reported significant traffic declines, as users found answers directly in the search results without clicking through to original sources. Some estimated traffic losses of 30-60% for pages that previously ranked well for informational queries. This created a perverse incentive: publishers had less motivation to create high-quality content if Google’s AI would simply scrape and repurpose it, potentially reducing their ad revenue and audience engagement.
The March 2024 Core Update: Volatility and Confusion
Despite the AI Overviews controversy, Google proceeded with its March 2024 core update, which it described as an effort to “address spam and low-quality results.” The update aimed to reduce unhelpful, unoriginal content by refining core ranking systems to better identify pages created for search engines rather than people. Google claimed the update would reduce low-quality content by 40%, later reporting an actual reduction of 45%. However, many publishers reported significant traffic volatility, with legitimate sites experiencing unexpected drops while low-quality AI-generated content initially thrived under the new quality-focused messaging.
The update’s implementation revealed complexities in Google’s ranking adjustments. Sites that had invested heavily in HCU-compliant, people-first content sometimes saw rankings fluctuate unpredictably, while thin AI-generated pages targeting specific long-tail queries occasionally gained traction. This inconsistency fueled skepticism about Google’s ability to consistently reward genuine quality. Webmasters reported spending considerable time diagnosing ranking changes, often finding no clear correlation between their improvements and search performance.
Adding to the confusion, the update interacted unpredictably with other Google systems. Sites using structured data or specific content management systems experienced disproportionate effects, suggesting that the update’s algorithms didn’t uniformly apply across different website architectures. For Malaysian publishers and businesses, this volatility meant that SEO strategies that worked reliably for years suddenly required constant adaptation, increasing the cost and complexity of maintaining online visibility.
Leadership Changes and Continued Struggles
By fall 2024, Google appeared to be struggling to regain control of search quality. In October 2024, the company replaced Prabhakar Raghavan as Head of Search amid mounting criticism. A November 2024 core update attempted to double down on the March update’s goals, pledging to surface “genuinely useful” content again. Yet the pattern continued: Google would crack down on low-value content, then seem to encourage it through AI-friendly policies, only to punish it again when the results proved unsatisfactory.
The leadership change reflected growing internal and external pressure to address search quality concerns. Raghavan’s tenure had seen numerous AI-focused initiatives, and his replacement signaled a potential shift in strategy. However, the November update’s mixed results suggested that deeper systemic issues persisted. Analysts noted that Google’s organizational structure—with semi-autonomous teams for Ads, AI, and Search—often led to conflicting priorities that undermined coherent long-term planning for search quality.
Throughout late 2024 and early 2025, users continued reporting frustrations with search results. Common complaints included excessive AI-generated content for simple queries, difficulty finding niche or local information amidst SEO-optimized pages, and a general sense that the search engine prioritized keeping users within Google’s ecosystem over directing them to the best external sources. These ongoing issues kept the “Google is getting weird” conversation alive in forums, social media, and tech publications.
AI Mode: The Next Evolution (or Misstep)?
In early 2025, Google unveiled AI Mode at its I/O event, presenting a chatbot-style search experience that attempted to answer queries without clearly citing sources. While positioned as a response to competing AI-powered search engines, AI Mode faced criticism for its lack of transparency and tendency to produce unattributed answers. Critics noted that Google had been beaten to market by more reliable AI search alternatives, raising questions about whether the company was innovating or merely playing catch-up in a rapidly evolving landscape.
AI Mode represented a more radical departure from traditional search than AI Overviews. Instead of providing summaries alongside organic results, AI Mode aimed to replace the conventional results page entirely with a conversational interface. Users would ask questions in natural language and receive direct answers, similar to interacting with a chatbot. However, early testing revealed significant shortcomings: answers often lacked proper attribution, making it difficult for users to verify sources or explore topics in depth. The system sometimes conflated information from multiple sources without clear distinctions, potentially spreading misinformation.
The launch also highlighted Google’s competitive disadvantages. Competitors like Perplexity and Arc Browser’s “Browse for Me” feature had already established reputations for more transparent, reliable AI-assisted search. These alternatives often provided clear source citations, balanced perspectives, and better handling of complex, nuanced queries. Google’s late entry into this space, combined with its ongoing trust issues, meant that AI Mode struggled to gain traction among users already skeptical of the company’s AI capabilities.
What This Means for Malaysian Users
For Malaysians who rely on Google for everything from daily news to research and local business information, these changes have tangible impacts. The increased prominence of AI-generated summaries means users must verify information more carefully, especially when seeking health, financial, or legal advice. Local businesses may find their visibility affected by algorithm shifts that favor certain types of content over others. Additionally, as competitors like Bing integrate AI more successfully and privacy-focused alternatives gain traction, Malaysian users have growing options beyond Google—though switching habits remains challenging due to habit and ecosystem integration.
In Malaysia, where smartphone penetration exceeds 85% and mobile data plans are relatively affordable, search behavior skews heavily toward mobile devices. This amplifies the impact of changes like increased ad density and AI Overviews, which consume valuable screen space on smaller displays. Malaysian users searching for local information—such as “best nasi lemak in Penang” or “COVID-19 vaccination centres Kuala Lumpur”—have reported encountering AI-generated answers that lack local context or accuracy, sometimes recommending businesses that have closed or providing outdated health guidelines.
Local businesses face particular challenges. The shift toward AI-generated content means that traditional SEO tactics focusing on keyword optimization may yield diminishing returns. Instead, businesses must focus on creating genuinely helpful, locally relevant content that stands a chance of being selected as a source for AI summaries or ranking well in organic results. This requires investment in quality content creation, local expertise, and user experience—resources that smaller enterprises may struggle to allocate.
Despite these challenges, Malaysian users are adapting. Surveys show growing awareness of search result limitations, with many users employing specific strategies: adding location modifiers to queries, using Malay language terms for locally relevant results, and cross-checking critical information across multiple sources. Some have begun exploring alternatives like DuckDuckGo for privacy-focused searching or Bing for its integration with Microsoft products. While Google remains the dominant search engine in Malaysia with over 90% market share, these adaptations suggest a maturing user base that is learning to navigate an increasingly complex search landscape.
Frequently Asked Questions
When did users first notice Google search getting worse?
While individual experiences vary, search volume data shows queries like “why is google search getting worse” began rising significantly in early 2022, coinciding with the Ads leadership shift and pre-dating the HCU launch. By mid-2023, related queries had increased fivefold compared to 2021 levels, indicating growing widespread concern.
Are AI Overviews available in Malaysia?
Yes, AI Overviews have been rolled out globally, including in Malaysia. Users may notice AI-generated summaries at the top of search results for certain queries, though they can be disabled by adding “-ai” to search queries. Malaysian users should be particularly cautious with health, financial, and legal queries, as AI errors in these areas could have serious consequences.
How can I improve my Google search results?
Experts recommend using more specific search terms, verifying information from multiple sources, trying alternative search engines like DuckDuckGo or Bing for comparison, and staying aware of confirmation bias when evaluating results. For Malaysian users, adding location-specific terms (like “Kuala Lumpur” or “Malaysia”) often improves local result relevance. Using quotation marks for exact phrases and the minus sign to exclude unwanted terms can also refine results.
Is Google still the best search engine for Malaysian users?
Google remains dominant in Malaysia with over 90% market share, but alternatives are improving. The “best” engine depends on individual needs—for quick answers, AI-powered alternatives may suffice, while for deep research, traditional search with careful verification still holds value. Users prioritizing privacy may prefer DuckDuckGo, while those invested in the Microsoft ecosystem might find Bing more convenient.
Will Google fix its search quality issues?
Google continues to release core updates and spam policy changes aimed at improving quality. However, the fundamental tension between advertising, AI ambitions, and user needs suggests that search quality will remain an ongoing challenge rather than a solved problem. The company’s quarterly earnings calls frequently reference “search quality” as a focus area, indicating ongoing awareness of the issue, but structural challenges persist.
How should Malaysian businesses adapt to these search changes?
Malaysian businesses should focus on creating genuinely helpful, locally relevant content that addresses specific user needs rather than chasing algorithmic trends. Investing in local language content, incorporating cultural nuances, and ensuring accurate, up-to-date information can improve visibility in both organic results and as potential sources for AI summaries. Monitoring search performance regularly and adapting strategies based on data—not just following generic SEO advice—is crucial for long-term success.
Conclusion
The perception that Google got “weird” didn’t stem from a single event but from a series of interconnected changes: advertising influences taking precedence, well-intentioned quality initiatives being reversed to favor AI ambitions, and rushed AI feature rollouts that prioritized innovation over reliability. For users worldwide, including Malaysians, the era of trusting Google search to simply “work” has given way to a more critical, verification-heavy approach. As Google continues to swing between cracking down on low-value content and embracing AI-generated material, users must adapt by staying informed, diversifying their information sources, and maintaining healthy skepticism toward any single search tool—no matter how dominant it once seemed.