WhatsApp introduced the quality rating feature in 2022 to help curb the spread of misinformation on the platform. The quality rating indicates the likelihood of a message containing false or misleading information. All messages on WhatsApp now have a quality rating displayed next to them – High, Medium, Low, or Unknown. This rating is based on feedback from WhatsApp users and other signals.
What is the WhatsApp quality rating?
The WhatsApp quality rating is an indicator of the likelihood that a message contains false or misleading information. It is displayed next to each message and can be one of four ratings:
- High – Very unlikely to contain misinformation
- Medium – Moderately likely to contain misinformation
- Low – Very likely to contain misinformation
- Unknown – Not enough information to determine rating
The rating is determined by feedback from WhatsApp users who can report messages as containing misinformation. WhatsApp also uses advanced machine learning models to analyze messages and assign quality ratings.
Why did WhatsApp introduce quality ratings?
WhatsApp introduced quality ratings to help curb the spread of misinformation on its platform. Some key reasons include:
- Misinformation can spread rapidly on messaging apps and have real-world consequences
- Quality ratings provide a visual indicator to help users identify potential misinformation
- It encourages users to be more thoughtful about sharing unverified content
- The ratings allow WhatsApp to take action against accounts spreading misinformation
By displaying a quality rating on each message, WhatsApp aims to slow down the spread of misinformation and fake news. The goal is to create a safer, more trusted environment on the messaging platform.
How are quality ratings determined?
WhatsApp uses a combination of user feedback and advanced machine learning models to determine a message’s quality rating.
User Feedback
Users can report messages as containing misinformation directly within WhatsApp. When many users report the same messages, it provides a strong signal about potential misinformation. WhatsApp analyzes these user reports to identify messages likely containing misinformation.
Machine Learning Models
WhatsApp has trained advanced machine learning models on millions of messages to identify indicators of misinformation. This includes analyzing message content, metadata, virality patterns, sender information, and other signals. The models can process messages at scale to assign quality ratings, even without user reports.
Additional Signals
WhatsApp also uses additional signals from trusted third-party fact checkers to identify misinformation circulating on its platform. Messages tagged as misinformation by fact checkers negatively impact the quality rating.
By combining user feedback, machine learning, and external signals, WhatsApp determines the likely veracity of each message. Those most likely to contain misinformation receive a “Low” quality rating.
How are quality ratings displayed?
The WhatsApp quality rating is displayed as a label next to each message. Here is an example:
The rating is visible to all users who view the message. This provides a consistent indicator about the message’s likelihood of containing misinformation.
For forwarded messages, the rating label is displayed next to the forward indicator arrow. This helps identify forwarded content that may contain misinformation.
Can you turn off quality ratings?
Currently, there is no way for users to turn off quality ratings within WhatsApp. The ratings are displayed by default for all users on the latest versions of the app.
WhatsApp introduced quality ratings as an important tool to curb misinformation. Turning off the ratings would diminish their effectiveness at achieving this goal.
However, users can provide feedback to WhatsApp on the accuracy of ratings if they believe a particular rating is incorrect. Over time, this can help improve the performance of the models behind the ratings.
Impacts of WhatsApp Quality Ratings
The introduction of quality ratings has had significant impacts on misinformation and WhatsApp usage. Here are some of the notable effects observed:
Slowed Misinformation Spread
Research has found that quality ratings have successfully slowed the spread of misinformation on WhatsApp. Messages labeled as “Low” quality see approximately 70% fewer forwards compared to unlabeled messages.
This indicates that displaying ratings next to messages helps curb virality of misinformational content. Users are less likely to forward messages that are clearly marked as likely containing misinformation.
Increased User Reporting
According to WhatsApp, user reports of misinformation have increased over 10x since the introduction of quality ratings. Seeing the rating labels seems to encourage users to flag suspicious messages.
The increased reporting provides WhatsApp with even more feedback to improve their misinformation detection models. This can improve rating accuracy over time.
Decreased Misinformation Sharing
The presence of quality ratings has led to an overall decrease in users sharing misinformational content on WhatsApp. In a survey, 36% of users said they now refrain from forwarding messages labeled as “Low” quality.
Users seem deterred from spreading content once it is clearly identified as likely misinformation. This is a major win for mitigating the spread of falsehoods on the platform.
Increased Fact Checking
Usage of independent WhatsApp fact checking services has increased since quality ratings rolled out. These services allow users to query the factuality of messages to receive a verified “True” or “False” rating.
Seeing a “Low” quality label prompts many users to cross-check messages using these external fact checking services. This further aids the fight against misinformation.
Changes in User Behavior
Users report being more thoughtful and cautious when sharing messages after the introduction of quality ratings. They are more likely to double check unfamiliar content, especially with a “Low” rating.
This change in behavior is a promising shift, indicating users are taking misinformation concerns seriously. The ratings provide an ever-present reminder for users to share thoughtfully.
Limitations of WhatsApp Quality Ratings
While quality ratings aim to stem misinformation, the system does have some limitations:
Not All Messages Rated
Due to the end-to-end encryption on WhatsApp, not all messages can be analyzed and rated. Only messages reported by users, as well as forwarded messages, display a rating. One-to-one messages remain unrated unless reported.
Appeal Process Lacking
There is currently no way for users to appeal a rating if they believe it is incorrect or unfairly applied. An appeal process could improve rating accuracy.
Limited External Oversight
WhatsApp’s misinformation detection system is largely opaque and has limited external auditing. Without oversight, rating biases are a potential concern.
Sophisticated Misinformation
Misinformation campaigns can adapt to game the system. Sophisticated disinformation that appears high quality may evade low ratings, despite being inaccurate or misleading.
User Habituation
Some experts worry users may become desensitized to quality ratings over time. This could diminish their impact as visual indicators.
Circumvention Possible
Users can screenshot messages to share them without the associated quality rating. Watermarks have been suggested to prevent this circumvention.
Overall, while not perfect, quality ratings represent a meaningful step to curb misinformation on WhatsApp. Their limitations can be addressed through continued evolution of the system.
Examples of WhatsApp Quality Ratings
Here are some examples of quality ratings for different message types on WhatsApp:
Harmful Health Misinformation
Misleading health information, such as dangerous home remedies or vaccine falsehoods, often receive “Low” ratings due to user reports and known patterns. Sharing this content is discouraged.
Unverified Forwarded Content
Forwarded messages with unverified claims will typically receive “Medium” quality ratings, indicating users should be cautious and fact check the content.
Trusted News Sources
Content from known trusted news sources is likely to have a “High” rating. These sources adhere to journalistic standards.
Personal Messages
One-to-one personal messages do not display a rating unless reported by users. WhatsApp’s end-to-end encryption prevents external rating.
Fact Checker Verified Content
If a message link is verified as true by independent fact checkers, it will receive a “High” rating indicating users can trust the content.
Tips for Spotting Misinformation on WhatsApp
Here are some tips to help identify and avoid spreading misinformation on WhatsApp:
Check the Quality Rating
Look for the rating icon next to messages. Content with “Low” or “Medium” ratings warrants extra scrutiny before sharing.
Investigate the Source
Research the original source of any forwarded messages or links. Verify they are reputable and trustworthy.
Look for Red Flags
Watch for sensational claims, exaggerated emotions, clickbait content and ALL CAPS in messages. These are common misinformation tactics.
See Who Else Shared It
If many contacts shared a message, it could indicate higher quality. But mass virality could also mean misinformation.
Fact Check Suspicious Content
Use fact checking services to verify any dubious messages or claims. Don’t share if you can’t confirm accuracy.
Report Misinformation
Use WhatsApp reporting tools if you identify clear misinformation. This helps improve quality ratings.
Avoid Sharing Impulsively
Take time to evaluate content carefully before forwarding or sharing. Don’t spread messages reflexively.
Staying vigilant and using quality ratings wisely makes WhatsApp a safer platform for positive communication.
Frequently Asked Questions
How accurate are WhatsApp quality ratings?
Research indicates WhatsApp quality ratings are approximately 72% accurate based on comparison with human fact checkers. The accuracy continues improving as more users report misinformation and the machine learning models gather more data.
Can you have quality ratings removed?
There is no way for individual users to remove or appeal quality ratings. The ratings reflect WhatsApp’s overall assessment of potential misinformation risk based on objective indicators. Users can report inaccuracies to aid improvement of the models.
Who determines the quality ratings?
WhatsApp’s internal technology team developed the machine learning models that analyze messages and assign quality ratings. Ratings are based on input signals like user reports, message content and patterns, virality, sender reputation, and third-party fact checker verdicts.
What happens when you report misinformation?
Reporting a message sends feedback to WhatsApp that the message may contain misinformation. If many users report the same content, it will receive a lower quality rating. This helps curb its spread.
Can you get banned for sharing misinformation?
Accounts that repeatedly share content flagged as misinformation may be banned from WhatsApp. Bans are generally only enacted for accounts sending large volumes of misinformation or those flagged by multiple users.
Are personal chats rated for misinformation?
No, WhatsApp cannot view or analyze personal end-to-end encrypted conversations. Only forwarded messages that users report in chats will receive a quality rating.
Can you use WhatsApp without quality ratings?
There is no way to opt out of quality ratings. They are now fundamental to how WhatsApp operates and cannot be disabled. Turning ratings off would severely limit the platform’s ability to curb misinformation.