A global review of nearly 400 prominent artificial intelligence accounts has found that online influence in the AI era is moving away from speed, volume and broad commentary toward verified technical capability, public testing and repeatable results.
The study, based on XHunt social data collected between March and May 2026, shows that generic AI information is losing value as audiences place more weight on creators who can demonstrate working software, publish transparent tests and explain how tools perform in real use. The findings suggest a clear change in how authority is formed online: visibility still matters, but measurable usefulness now matters more.
The shift is especially important for technology markets, digital asset traders and financial professionals who have long relied on social platforms for fast information. The data indicate that follower counts and simple view totals are becoming weaker signals of practical value. Instead, social authority is increasingly tied to whether a person can automate complex tasks, test systems in public and deliver results that others can reproduce.
The report also highlights a sharp divide between English-language and Chinese-language AI communities. English-language AI figures continue to dominate in total reach, backed by older accounts, established research figures and company leaders. Chinese-language creators, by contrast, post more often and focus more heavily on tutorials, product testing, coding workflows and AI agents used in day-to-day operations.
The result is a two-track global AI influence system. One side is led by institutional voices and deep technical research. The other is driven by practical deployment, fast experimentation and community learning.
Influence now depends on proof
The main conclusion of the study is that influence in AI is no longer measured mainly by how quickly someone reacts to news or how often they post. The more valuable signal is sustained public verification.
That means social accounts gain trust when they show how an AI tool works, publish tests, compare outcomes and explain limitations. In earlier internet cycles, rapid information sharing often helped accounts grow. In the generative AI cycle, audiences appear to be rewarding creators who can turn information into usable methods.
This is a major change for online technology discussion. Many AI users are no longer satisfied with broad predictions or recycled news. They want examples, code, benchmarks, workflows and evidence that a tool is useful in real settings.
The report found that this trend became clearer over the three-month study period. In March 2026, total AI-related posts among the tracked accounts rose to 12,400 and generated 310 million views. But average engagement fell to about 25,000 per post, partly because many accounts repeated the same topics and covered the same model releases.
By May, the number of posts declined to 9,000. Yet total views climbed to 335 million, and average views per post rose to about 37,400. The study interpreted that change as a sign that audiences were filtering out repeated content and spending more attention on technically useful posts.
In simple terms, fewer posts produced better results when those posts were more practical, more original or more clearly tested.
English-language accounts lead in reach
English-language AI accounts remain much larger by audience size. According to the study, the top 300 English-language accounts had a combined follower count of more than 350 million. That equals an average of about 1.17 million followers per account.
The English-speaking group also has a longer history online. About 62.9% of influential English-language accounts were registered between 2007 and 2015. Many of these accounts belong to people who built influence across earlier internet and software cycles before generative AI became mainstream.
That older base gives the English-language AI community a structural advantage. Many of its leading voices are already known in machine learning, software engineering, venture-backed technology or academic research. Their authority did not appear suddenly during the AI boom. It was often built over many years.
The study classified more than 65% of leading English-language accounts as company founders, executives or scientists. These figures often use social platforms to distribute research, explain product direction, comment on model architecture or shape broader technical debate.
Among the highest-ranked English-language figures were Andrej Karpathy, Sam Altman and Jeff Dean. The report described them as examples of institutional and research leadership, with influence supported by long-term technical credibility and ties to major AI organizations.
English-language AI figures also scored strongly in areas linked to model architecture and multimodal systems. The report gave them an average technical score of 88.3 in those categories. That reflects a focus on core AI design, model behavior, training methods and the combination of text, image, audio and other data types.
This gives the English-language ecosystem an important role in shaping high-level AI narratives. It is where much of the discussion around frontier models, research direction and large-scale product strategy takes place.
Chinese-language creators post more often
The Chinese-language AI community is smaller in audience size but more active in daily posting. The top 100 Chinese-language accounts averaged about 77,000 followers, far below the average reach of the English-language group.
However, Chinese-language creators posted far more frequently. They averaged 210 posts over the 90-day study period, compared with 38 posts among English-language counterparts. This created a dense information network built around daily tool use, testing and workflow experimentation.
The study found that 13% of influential Chinese-language AI accounts emerged between 2022 and 2023. That timing reflects the rapid growth of content tied to generative AI applications, tutorials and hands-on product use after the public release of powerful chatbot and coding tools.
Rather than focusing mainly on model design or institutional strategy, many Chinese-language creators emphasize how people can use AI tools immediately. Their content often includes coding demonstrations, agent-building examples, prompt workflows, automation templates and comparisons of different products.
The report classified 69.1% of leading Chinese-language AI accounts as tool evaluators or product engineers. That is a very different profile from the English-language group, where founders, executives and scientists were more common.
This practical orientation was also reflected in technical scoring. Chinese-language figures scored 88.9 in AI programming and 87.1 in agent-based applications. The report said this showed a stronger tilt toward full-stack integration, coding workflows and real-world deployment.
Several Chinese-language accounts led this practical community, including Bao Yu, Orange AI and Gui Zang. The study described them as important nodes in open tutorials, transparent coding sessions and AI agent deployment experiments.
Different tools, different adoption patterns
The study also tracked the tools and models most often mentioned by major Chinese-language accounts. Claude was mentioned by 95.7% of major Chinese accounts, while Codex was referenced by 80.9%.
Domestic AI models also showed strong adoption. DeepSeek was mentioned by 68.1% of the tracked Chinese-language accounts, and Kimi was mentioned by 58.5%.
These figures suggest that Chinese-language AI creators are not focused only on global tools or only on domestic systems. Instead, they appear to combine both, choosing products based on coding performance, workflow value and availability.
This blended pattern is important because it shows how AI adoption spreads through practical communities. Tools gain traction when creators can show clear use cases, not just when companies announce new features.
In many cases, the most influential content is not a product announcement. It is a detailed demonstration showing how a model writes code, controls an agent, searches documents, automates a business process or connects with other software.
That type of content is harder to fake than general commentary. It can be tested by others. It can be challenged. It can fail in public. Those qualities make it more valuable as a trust signal.
Audiences favor analytical voices
The report also included personality-style analysis of leading accounts. Most influential AI figures were placed within the rational “NT” spectrum, a category associated with analytical, systematic and instructional communication.
The study said this suggests that audiences prefer clear reasoning and technical explanation during periods of rapid technological change. When tools are evolving quickly, people appear to reward accounts that reduce confusion and explain how systems work.
This is another sign that AI influence differs from older social media influence models. Personality and entertainment still matter, but in technical communities, practical clarity carries more weight.
In the current AI cycle, the strongest accounts are not simply loud or fast. They are useful. They help audiences understand what to do with new tools.
AI and Web3 influence follow different rules
The report also compared AI commentators with Web3 influencers and found different value structures.
Web3 networks are often built around the distribution of information, market opportunities and sentiment. Speed has historically been valuable in those communities because prices can move quickly and attention can shift from one asset narrative to another in hours.
AI networks, by contrast, increasingly trade in methods. Their value comes from reproducible problem-solving, workflow design and technical efficiency. A strong AI post often teaches someone how to build something, automate a task or improve a process.
That distinction matters as the two sectors begin to overlap. The report noted that some former crypto analysts are now using AI agents for automation, research and market monitoring. Digital asset traders are also exploring AI tools to test strategies, filter information and manage risk.
But the study’s findings suggest that traders should be careful about relying on social sentiment alone. In a faster and more automated market environment, random posts and rumor-driven narratives may be less useful than tested systems and repeatable processes.
Automation is changing finance
The broader financial sector is already moving in the direction described by the AI influence study. Recent industry numbers show that computer algorithms handle roughly 70% of daily trading volume across major equity markets.
That level of automation means human speed alone is no longer enough in many market settings. Traders competing with automated systems need better tools, better rules and better testing. Reacting to social media posts after they spread widely is unlikely to offer a durable edge.
The use of smart software is also spreading across companies. One recent study found that 88% of modern firms use smart tools in daily operations. However, fewer than 10% have successfully scaled those automated programs to produce clear returns.
That gap is important. It shows that adopting AI tools is not the same as using them well. Many organizations can test chatbots, automate simple tasks or buy AI software. Far fewer can build reliable systems that improve decisions, lower costs or generate measurable results.
Corporate leaders are preparing for broader use. Research cited in the report shows that 61% of chief executives are actively preparing to use automated agents at large scale this year.
Automated agents are software systems that can carry out tasks with limited human direction. In finance, they can monitor markets, summarize documents, test strategies, track risk conditions and execute predefined workflows. In business operations, they can handle customer requests, manage data, prepare reports and coordinate software tools.
Aral, cited in the research, said every operating entity needs a clear strategy to deploy and use these agents properly. That warning reflects a growing concern in the market: automation can improve productivity, but poorly designed systems can also create errors, false confidence and new risks.
The market value of followers is weakening
The collected data indicate that follower counts and simple views are no longer reliable measures of market value or professional relevance. Large audiences can still amplify a message, but reach alone does not prove competence.
In AI, the more important measure is whether a creator can show working results. This includes public tests, transparent failures, detailed tutorials and tools that others can use. In financial markets, the equivalent is whether a strategy can be tested, repeated and adjusted under changing conditions.
For digital asset traders, the lesson is direct. The market is becoming more dependent on automation, data processing and fast execution. Chasing random online rumors or daily social sentiment is increasingly risky when algorithms dominate liquidity and when AI-generated content can flood platforms with low-quality information.
A more durable approach is to use smart software to test plans, monitor risk and compare outcomes. That does not require every trader to become a professional engineer. But it does require more technical discipline than simply following viral accounts.
The new influence model rewards people who can build, test and explain. The same standard is spreading into markets.
Reliability is becoming the new authority
The study’s final conclusion is that online authority in the AI era is shifting from simple visibility to measurable reliability.
That shift may reshape how traders, companies and technology communities evaluate expertise. The most trusted voices are increasingly those who prove their claims in public and provide tools or methods that work beyond a single post.
English-language AI communities still command the largest audiences and remain central to research-led discussion. Chinese-language communities are building influence through high-frequency experimentation, product testing and practical deployment. Both ecosystems point toward the same larger trend: AI influence is becoming more evidence-based.
For markets, the message is clear. Speed still matters, but proof matters more. The accounts, firms and traders that adapt fastest to that reality are likely to gain the most from the next stage of automation.
For deeper insight into AI‑driven trading shifts, explore our guide on AI copy trading in modern crypto markets.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
