There is not always a person behind an online conversation. Comments, shares, likes, and entire discussions can be produced or amplified by online bots, software designed to perform certain activities on the web automatically.
Prenota ora un'analisi gratuita della tua reputazione
Bots are not necessarily harmful: they are used for customer service, indexing, notifications, and moderation. The problem arises when automated accounts and coordinated networks are used to simulate consensus, artificially increase the visibility of a narrative, or influence perceptions of people and companies.
Determining how many bots actually exist is complex. There is no universal percentage that applies to every platform. From a reputational perspective, however, the central issue is different: even a limited number of coordinated accounts can generate enough interactions to make an opinion appear dominant when it is not.
What are online bots?
An online bot is a program capable of automatically performing activities that could otherwise be carried out by a person.
In the context of social networks, people often refer to social bots: automated or partially automated profiles that can publish content, follow accounts, share posts, like, or comment.
Automation alone does not indicate manipulative activity. The difference lies in the bot’s purpose and the transparency with which it is used.
A company chatbot that is clearly identified as automated performs an evident function. By contrast, a network of profiles presented as real users and coordinated to create artificial consensus or dissent poses a reputational problem.
How many bots are really online?
Determining what percentage of social media accounts are bots is difficult. Platforms use different criteria, and some profiles combine human and automated activity.
For this reason, there is no single, universally recognized percentage of bots on social media.
Significant estimates nevertheless circulate in public debate. During the 2025 World Economic Forum in Davos, Spanish Prime Minister Pedro Sánchez cited data suggesting that around one third of social media profiles may be bots and that almost half of Internet traffic may be generated by them. These estimates were cited in a political speech and do not constitute a universal measurement of every platform, so they should be interpreted with caution.
The figure that matters for reputation is another one: the impact of an automated network depends not only on the number of accounts but also on its ability to coordinate and amplify a message.
Online reputation monitoring should therefore not be limited to measuring how many people talk about a brand; it should also analyze the origin and dynamics of the conversation.
How social bots can alter a conversation
An automated network does not necessarily need to persuade thousands of people directly. It may be enough to change how popular a piece of content appears.
If many accounts publish the same message simultaneously, promote a hashtag, or repeatedly criticize a company, a real user may perceive the conversation as far more widespread than it actually is.
This is the principle of fake engagement: apparently organic interactions that do not necessarily reflect genuine public interest.
Bots can be used to:
- artificially increase likes, comments, and shares;
- make content or hashtags more visible;
- repeat the same accusation continuously;
- create the impression of widespread consensus or dissent;
- amplify false or context-free information.
The result is a distortion of collective perception. Content may appear relevant because it seems to involve a large number of people, even though some of the interactions were generated artificially.
Fake engagement and online reputation
The number of comments, volume of conversations, and frequency of mentions all contribute to public perceptions of an issue. When these indicators are manipulated, the representation of reality can also be altered.
Imagine that a company suddenly receives hundreds of negative comments. It might look like the beginning of a crisis caused by dissatisfied customers. A deeper analysis, however, might reveal recently created profiles, nearly identical messages, and interactions concentrated within the same time intervals.
In this case, the problem is not only the negative content but the artificial construction of a negative perception.
To distinguish genuine protest from a possible coordinated campaign, reputational risk analysis becomes essential for assessing the origin, spread, and characteristics of the accounts involved.
Bots and misinformation: why automation increases reach
Bots become particularly problematic when they are used to distribute misinformation.
The World Economic Forum’s Global Risks Report 2026 ranks misinformation and disinformation second among global risks over the next two years. The report also highlights how increasingly difficult it is to distinguish authentic from synthetic online content in an information ecosystem transformed by social media and artificial intelligence.
Automation can amplify the phenomenon even further. A false story may be published only once; a network of accounts can instead repost and comment on it repeatedly, creating an appearance of relevance that may encourage its spread among real users.
This is one of the clearest points of contact between fake news and online reputation: even false content can have concrete consequences when it is perceived as credible and reaches a large audience.
How to recognize a possible bot network
Not every spike in unusual activity represents an attack. The accounts’ overall behavior must be examined.
The most significant warning signs include:
- a sudden increase in mentions;
- repetition of the same language;
- accounts created during the same period;
- an unusually high posting frequency;
- interactions focused on a single topic;
- coordinated sequences of reposts, likes, or comments.
No single element automatically proves that an account is a bot. It is the combination of behaviors that makes a conversation worthy of further investigation.
Automated accounts and real users may also act at the same time. An artificial network may start a conversation that is subsequently picked up by real people, making the origin of the manipulation progressively less evident.
Bots and artificial intelligence: more credible automated accounts
Generative artificial intelligence is also making automation more sophisticated.
In the past, many bots were relatively easy to identify because they posted identical messages or displayed extremely regular patterns. Today, generative systems can produce different texts, adapt language to the context, and make automated interactions more plausible.
In 2025, the World Economic Forum highlighted how AI, deepfakes, and bots can make misinformation faster, more convincing, and harder to detect, with potential economic and reputational consequences for organizations.
It is therefore becoming less effective simply to ask whether a profile “looks fake.” What matters is analyzing how it behaves within the network and what relationships it has with other accounts.
When bots become a reputational crisis
An automated campaign can make a minority position appear dominant, increase the visibility of a marginal news story, or artificially amplify a controversy.
The risk increases when the narrative moves beyond the initial network of accounts and is picked up by real users, communities, influencers, or the media. At that point, an activity that began artificially can produce real reputational consequences.
For this reason, reputational crisis management must begin with an understanding of the dynamics, not merely the number of visible negative comments. A disproportionate public response, especially in the early stages, could help increase the campaign’s visibility.
What to do if you suspect a bot campaign
The first reaction should not be to respond to every suspicious profile. The size, origin, and impact of the phenomenon must be understood first.
An effective strategy should include:
- Preserve the evidence, including URLs, screenshots, accounts, and times.
- Analyze the patterns, looking for similarities in messages and behavior.
- Measure the spread, distinguishing suspicious accounts from real users.
- Assess the reputational impact, checking whether the narrative reaches relevant media, search engines, or communities.
- Define a proportionate response, choosing among monitoring, reporting, and public communication.
Social media monitoring tools make it possible to observe how conversations evolve and identify anomalies that are difficult to recognize by manually analyzing individual comments.
Reputation is not measured by counting interactions alone
Likes, comments, followers, and shares are visible indicators, but they do not always represent real people, consensus, or genuine opinions.
This is the central point in the relationship between bots and online reputation. A large volume of conversation does not necessarily correspond to a large number of people.
For this reason, digital reputation management requires a qualitative as well as quantitative reading of the data.
It is not enough to know how much a brand is being discussed. You need to understand who is talking about it, why, and through what amplification dynamics.
Distinguishing an authentic conversation from one constructed artificially is now an essential component of protecting the reputation of people and organizations.
Frequently asked questions about online bots
Online bots are programs that automatically perform certain activities on the web. They may have legitimate functions or be used to generate artificial interactions and amplify content.
There is no universal percentage that applies to all platforms. Estimates vary according to the social network, the period analyzed, and the criteria used.
Unusual posting frequency, repetitive messages, and coordinated behavior can be useful warning signs. However, no single indicator can identify a bot with certainty.
Yes. Automated networks can amplify criticism, fake news, or negative content and create an artificial perception of consensus or dissent, influencing real users as well.
The suspicious activity must be documented, the accounts involved analyzed, the spread of the narrative monitored, and its impact assessed before deciding how to respond.
