Media Manipulation and Bias Detection
Auto-Improving with AI and User Feedback
HonestyMeter - AI powered bias detection
CLICK ANY SECTION TO GIVE FEEDBACK, IMPROVE THE REPORT, SHAPE A FAIRER WORLD!
Government of Jamaica / official data
Caution! Due to inherent human biases, it may seem that reports on articles aligning with our views are crafted by opponents. Conversely, reports about articles that contradict our beliefs might seem to be authored by allies. However, such perceptions are likely to be incorrect. These impressions can be caused by the fact that in both scenarios, articles are subjected to critical evaluation. This report is the product of an AI model that is significantly less biased than human analyses and has been explicitly instructed to strictly maintain 100% neutrality.
Nevertheless, HonestyMeter is in the experimental stage and is continuously improving through user feedback. If the report seems inaccurate, we encourage you to submit feedback , helping us enhance the accuracy and reliability of HonestyMeter and contributing to media transparency.
Using wording or framing that makes a situation seem more dramatic or alarming than the underlying data alone would suggest.
Headline and subhead: "BPO’S US$220-M SLIDE – Industry sheds 5,000 jobs as local spending falls to US$780 million". Body text is measured and technical, but the word "SLIDE" in all caps, paired with the job-loss framing, can prime readers to perceive a more catastrophic collapse than the article’s detailed context (e.g., comparison with other exports, explanation of causes, policy responses) actually supports.
Use more neutral headline wording that still conveys the decline but avoids dramatic connotations, for example: "BPO sector records US$220m decline in local spending" or "BPO local spending falls to US$780m amid 5,000 job losses".
Avoid all-caps styling for the word "SLIDE" in contexts where it can be read as shouting or exaggerating the severity of the change.
Balance the subhead by briefly signalling both the decline and the policy/adjustment response, e.g., "Industry sheds 5,000 jobs as local spending falls to US$780 million; working group targets higher-value services".
Using emotionally charged or symbolic imagery or descriptions to influence readers’ feelings rather than focusing strictly on neutral presentation of facts.
Caption: "An unused headset at a vacant workstation symbolises the contraction in Jamaica’s BPO sector, which shed 5,000 jobs as annual local spending fell by US$220 million. (AI-generated photo illustration)." The image and the word "symbolises" are explicitly chosen to evoke a sense of abandonment and loss. While not extreme, this is a mild emotional framing layered on top of already clear quantitative data.
Rephrase the caption in more neutral, descriptive terms, for example: "An AI-generated illustration of a BPO workstation. The sector shed 5,000 jobs as annual local spending fell by US$220 million."
Avoid attributing symbolic meaning in the caption ("symbolises the contraction") and instead let readers draw their own conclusions from the data presented in the article.
If possible, pair the image with a second, more forward-looking visual (e.g., training or technology upgrades) to balance the emotional framing and reflect both challenges and responses.
Leaving out clarifying details that would help readers fully understand discrepancies or context, even if not done with clear bias.
On employment discrepancies: "However, Global Services Association of Jamaica President-elect Yoni Epstein has placed employment in the sector at about 40,000 for two consecutive years... The Government, by comparison, reported approximately 55,000 jobs in March 2025 and 50,000 in March 2026, leaving differences of 15,000 and 10,000, respectively. It was not immediately clear whether the discrepancies reflected different reporting periods, industry coverage or methods of counting workers." The article notes the discrepancy and admits uncertainty, which is good practice, but it stops there. Readers are left without any exploration of likely methodological differences (e.g., inclusion of part-time, contractors, or related services), which would be important context for interpreting the numbers.
Add a brief explanatory paragraph outlining the most plausible reasons for the discrepancy, even if approximate, for example: "Industry sources suggest that government figures may include related support services and part-time workers, while GSAJ’s estimate may focus on core BPO roles only."
Explicitly state whether the reporter sought clarification from either the Government or GSAJ on the counting methods and whether any response was received.
If no clarification is available, clearly flag this as a limitation: "Without detailed methodological notes from either source, it is not possible to reconcile the figures at this time."
Providing more space or detail to one side’s perspective or data than to others, which can subtly favor that side even without overt bias.
The article relies heavily on the Government’s annual reports and official data (Form 18-K, DBJ, PAJ, Jampro, JSEZA) and gives them detailed exposition. The industry side is represented mainly through a single figure (Yoni Epstein’s 40,000 jobs estimate) and a brief mention of GSAJ, without deeper exploration of industry critiques, alternative explanations, or on-the-ground impacts. This is not overtly biased, but structurally it gives more narrative and explanatory weight to official perspectives than to industry or worker perspectives.
Include at least one additional industry voice (e.g., another BPO operator, an industry analyst) to comment on the causes of the contraction and the adequacy of the Government’s response.
Add a short section summarising how workers or unions view the job losses and the shift toward KPO, if such perspectives are available.
Explicitly note the reliance on official documents as a limitation, for example: "This article is based primarily on Government filings and agency reports; industry and worker perspectives were not extensively available at press time."
- This is an EXPERIMENTAL DEMO version that is not intended to be used for any other purpose than to showcase the technology's potential. We are in the process of developing more sophisticated algorithms to significantly enhance the reliability and consistency of evaluations. Nevertheless, even in its current state, HonestyMeter frequently offers valuable insights that are challenging for humans to detect.