{"id":2143532196,"date":"2026-09-25T16:42:19","date_gmt":"2026-09-25T11:12:19","guid":{"rendered":"https:\/\/newgensoft.com\/resources\/?p=2143532196"},"modified":"2026-09-25T16:42:53","modified_gmt":"2026-09-25T11:12:53","slug":"what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong","status":"publish","type":"post","link":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/","title":{"rendered":"What Happens When AI Explains Itself Beautifully, and Gets It Wrong"},"content":{"rendered":"\n

For insurers, explainability is only the beginning. The real test is whether people can still question the machine.<\/em><\/strong><\/strong><\/p>\n\n\n\n

There is a question insurance leaders should start asking about AI.<\/p>\n\n\n\n

Not simply: Can the AI explain its decision?<\/strong><\/p>\n\n\n\n

But: What happens when the explanation sounds perfectly reasonable and is still wrong?<\/strong><\/p>\n\n\n\n

In an AI-supported underwriting process, a recommendation may come with everything we would expect from a responsible system: the decision, the factors behind it, a confidence score and a human reviewer who can approve or challenge it.<\/p>\n\n\n\n

The difficult part is that the explanation itself can still be inaccurate.<\/strong><\/p>\n\n\n\n

A 2026 study on explanatory misinformation<\/a> in human-AI collaboration examined situations in which AI produced the correct answer while the explanation supporting that answer was incorrect. People exposed to incorrect explanations later showed weaker reasoning than those who received AI advice without the faulty explanation. In effect, the system could help someone reach the right answer while leaving behind the wrong understanding of why that answer was right.<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

For insurance, the implication goes beyond a single decision. An AI-generated explanation can influence how an underwriter, claims professional or fraud investigator approaches the next case as well<\/strong>.<\/p>\n\n\n\n

A good explanation is not automatically a good reason<\/strong><\/h2>\n\n\n\n

The original concern around explainable AI was straightforward, if an algorithm influences an important decision, the person using it should be able to understand why.<\/p>\n\n\n\n

That remains important. What\u2019s becoming equally important is the quality of the explanation itself<\/strong>.<\/p>\n\n\n\n

An AI might state that \u2018This claim deserves additional investigation because A, B and C indicate possible fraud.\u2019<\/strong><\/p>\n\n\n\n

The important question is not how clearly A, B and C have been presented. It is whether those factors genuinely support the conclusion and whether the reviewer can verify them.<\/p>\n\n\n\n

That means \u2018explanation available\u2019 cannot by itself be treated as evidence that human oversight is working.<\/p>\n\n\n\n

A 2023 study<\/a> comparing seven different forms of AI explanation tested approaches including probability scores, feature importance, decision trees, natural-language explanations, counterfactuals and case-based reasoning. None of the explanation formats significantly changed how often participants followed the AI\u2019s advice, including when the advice was incorrect. Probability scores were also rated significantly less explainable than several other approaches.<\/p>\n\n\n\n

Explaining a mistake does not stop it from being a mistake.<\/strong><\/p>\n\n\n\n

Explainability is useful when it helps a person understand and examine a recommendation, not simply when it makes the recommendation easier to accept.<\/p>\n\n\n\n

AI can improve individual decisions while making decisions more alike.<\/strong><\/h2>\n\n\n\n

This may be an even more interesting question for insurers. A 2021 MIS Quarterly study<\/a> on humans working with AI advice found that individual accuracy increased from 68.1% without AI advice to 79.9% with it<\/strong>.<\/p>\n\n\n\n

At the same time, what the researchers called unique human knowledge<\/strong>, situations in which the human was correct while AI was wrong, fell from 12.3% to 7.3%<\/strong>. The experiments were based on image-classification tasks rather than insurance, so these figures should not be treated as underwriting benchmarks. The underlying behaviour, however, raises an important organisational question.<\/p>\n\n\n\n

Experienced underwriting teams do not add value only through consistency. Their value also lies in differences of judgment. One underwriter may notice an unusual pattern in a medical history. Another may connect the case with something seen years earlier. Another may question whether two apparently related factors should really be assessed together. Those differences can sometimes slow a discussion down, but they are also part of what makes expert judgment valuable.<\/p>\n\n\n\n

The same MIS Quarterly research<\/strong><\/a> <\/strong>found that AI advice caused human choices to converge. Individuals became more accurate, but they also became more likely to make similar choices and therefore share the same AI blind spots. This can weaken the \u2018wisdom of crowds\u2019<\/strong> when multiple people are influenced by the same machine advice.  For insurance leaders, that creates a useful question:<\/p>\n\n\n\n

Are we using AI to strengthen expertise, or could greater consistency eventually reduce some of the independent judgment that makes expertise valuable?<\/strong><\/p>\n\n\n\n

Perhaps the human should sometimes form a view first.<\/strong> The point at which AI is introduced into a decision can significantly shape the outcome.<\/strong><\/p>\n\n\n\n

In a 2023 MIS Quarterly study<\/strong><\/a> of AI-supported diagnostic work, physicians made their own diagnosis before seeing the AI-supported assessment. The study found that doctors made their own diagnosis before seeing the AI’s recommendation. This helped them think independently first.<\/p>\n\n\n\n

Many AI-supported processes effectively begin with the machine recommendation and then ask a human to review it. For high-volume, straightforward work, that can make perfect sense. For some complex decisions, another sequence may be worth considering i.e., allow the expert to reach an initial view first, then use AI as an additional perspective and investigate where the two differ.<\/p>\n\n\n\n

That changes AI from being the first voice in the decision<\/strong> to being a second pair of eyes<\/strong>. For complex medical underwriting, unusual claims, fraud investigations or cases with significant customer impact, preserving that independent first view may be valuable alongside improvements in speed and consistency.<\/p>\n\n\n\n

It also means a very high human-AI agreement rate needs to be interpreted carefully. High agreement may reflect strong model performance. But the diagnostic-work research<\/strong><\/a> <\/strong>also showed why agreement itself should not be treated as proof of better thinking, agreement and disagreement can produce very different levels of reflection.<\/p>\n\n\n\n

Confidence needs context<\/strong><\/p>\n\n\n\n

The same caution applies to AI confidence.<\/p>\n\n\n\n

A 2024 Journal of the Association for Information Systems paper<\/a> on human-in-the-loop AI reviewing used GPT-4 in an academic-reviewing workflow based on 2,040 papers and 7,698 human reviews<\/strong>. GPT-4 reported high confidence in more than 80% of cases, while human reviewers showed a much wider spread of confidence levels. The work also identified bias and showed how adding relevant contextual information could materially change the model\u2019s behaviour.<\/p>\n\n\n\n

This was not an insurance experiment, and its results should not be transferred directly to underwriting. The principle is more useful than the percentages:<\/p>\n\n\n\n

AI confidence is not the same as evidence.<\/strong><\/p>\n\n\n\n

If an underwriting copilot displays a high-confidence recommendation, the number should begin a conversation rather than end one.<\/p>\n\n\n\n

The useful questions are where the recommendation came from, which information the model used, what information was missing, what could change the outcome, whether the recommendation aligns with policy and risk rules, and whether the underwriter has enough information to disagree meaningfully.<\/p>\n\n\n\n

So what should insurance leaders ask?<\/strong><\/h2>\n\n\n\n
    \n
  1. Can we verify the AI\u2019s explanation, not merely display it?<\/li>\n\n\n\n
  2. For complex cases, are professionals able to form an independent view before AI begins influencing the decision?<\/li>\n\n\n\n
  3. When people override AI, are we examining the quality of their reasoning rather than simply counting overrides?<\/li>\n\n\n\n
  4. Are multiple reviewers becoming dependent on the same model and therefore becoming exposed to the same blind spots?<\/li>\n\n\n\n
  5. Are there situations where AI advice should be delayed or withheld because an independent human view adds more value?<\/li>\n<\/ol>\n\n\n\n

    The real goal is not simply explainable AI<\/strong><\/h2>\n\n\n\n

    Insurance does not need AI that produces an answer and then adds a paragraph underneath it. There\u2019s a dire need for AI whose recommendations can be understood and checked, whose limitations can be recognized, and whose role changes according to the complexity of the decision.<\/p>\n\n\n\n

    It also needs professionals whose expertise grows alongside AI rather than becoming dependent on it.<\/strong><\/p>\n\n\n\n

    That is why the conversation should move beyond \u2018Is our AI explainable?\u2019<\/strong><\/p>\n\n\n\n

    A more useful question is \u2018After years of working with AI, will our people still know when it is wrong?\u2019<\/strong> This may ultimately be the difference between putting a human somewhere in the process and keeping human judgment genuinely part of the process.<\/strong><\/p>\n","protected":false},"excerpt":{"rendered":"

    For insurers, explainability is only the beginning. The real test is whether people can still question the machine. There is a question insurance leaders should start asking about AI. Not simply: Can the AI explain its decision? But: What happens when the explanation sounds perfectly reasonable and is still wrong? In an AI-supported underwriting process, […]<\/p>\n","protected":false},"author":6,"featured_media":2143532206,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1],"tags":[],"resources-topic":[],"resources-group":[],"class_list":["post-2143532196","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blog"],"acf":[],"yoast_head":"\nWhat Happens When AI Explains Itself Beautifully, and Gets It Wrong - Newgen Resources (DX++)<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"What Happens When AI Explains Itself Beautifully, and Gets It Wrong - Newgen Resources (DX++)\" \/>\n<meta property=\"og:description\" content=\"For insurers, explainability is only the beginning. The real test is whether people can still question the machine. There is a question insurance leaders should start asking about AI. Not simply: Can the AI explain its decision? But: What happens when the explanation sounds perfectly reasonable and is still wrong? In an AI-supported underwriting process, […]\" \/>\n<meta property=\"og:url\" content=\"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/\" \/>\n<meta property=\"og:site_name\" content=\"Newgen Resources (DX++)\" \/>\n<meta property=\"article:published_time\" content=\"2026-09-25T11:12:19+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-09-25T11:12:53+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/newgensoft.com\/resources\/wp-content\/uploads\/sites\/6\/2026\/09\/For-Blog.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1536\" \/>\n\t<meta property=\"og:image:height\" content=\"1024\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Darshak Parmar\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Darshak Parmar\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"6 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/\"},\"author\":{\"name\":\"Darshak Parmar\",\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/#\\\/schema\\\/person\\\/03d87995d5366558c449b58979bdd031\"},\"headline\":\"What Happens When AI Explains Itself Beautifully, and Gets It Wrong\",\"datePublished\":\"2026-09-25T11:12:19+00:00\",\"dateModified\":\"2026-09-25T11:12:53+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/\"},\"wordCount\":1257,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/wp-content\\\/uploads\\\/sites\\\/6\\\/2026\\\/09\\\/For-Blog.jpg\",\"articleSection\":[\"Blog\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/\",\"url\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/\",\"name\":\"What Happens When AI Explains Itself Beautifully, and Gets It Wrong - Newgen Resources (DX++)\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/wp-content\\\/uploads\\\/sites\\\/6\\\/2026\\\/09\\\/For-Blog.jpg\",\"datePublished\":\"2026-09-25T11:12:19+00:00\",\"dateModified\":\"2026-09-25T11:12:53+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/#primaryimage\",\"url\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/wp-content\\\/uploads\\\/sites\\\/6\\\/2026\\\/09\\\/For-Blog.jpg\",\"contentUrl\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/wp-content\\\/uploads\\\/sites\\\/6\\\/2026\\\/09\\\/For-Blog.jpg\",\"width\":1536,\"height\":1024},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/blog\\\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"What Happens When AI Explains Itself Beautifully, and Gets It Wrong\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/#website\",\"url\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/\",\"name\":\"Newgen DX++\",\"description\":\"\",\"publisher\":{\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/#organization\",\"name\":\"Newgen Software\",\"url\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/wp-content\\\/uploads\\\/sites\\\/6\\\/2026\\\/06\\\/Newgen-Logo_Light-scaled.png\",\"contentUrl\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/wp-content\\\/uploads\\\/sites\\\/6\\\/2026\\\/06\\\/Newgen-Logo_Light-scaled.png\",\"width\":2560,\"height\":788,\"caption\":\"Newgen Software\"},\"image\":{\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/#\\\/schema\\\/logo\\\/image\\\/\"}},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/newgensoft.com\\\/resources\\\/#\\\/schema\\\/person\\\/03d87995d5366558c449b58979bdd031\",\"name\":\"Darshak Parmar\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/ded91f8c09d9e3c95bc58254ae8fa31f63e4d843b2bbe2bbbe15d5254bd1778b?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/ded91f8c09d9e3c95bc58254ae8fa31f63e4d843b2bbe2bbbe15d5254bd1778b?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/ded91f8c09d9e3c95bc58254ae8fa31f63e4d843b2bbe2bbbe15d5254bd1778b?s=96&d=mm&r=g\",\"caption\":\"Darshak Parmar\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"What Happens When AI Explains Itself Beautifully, and Gets It Wrong - Newgen Resources (DX++)","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/","og_locale":"en_US","og_type":"article","og_title":"What Happens When AI Explains Itself Beautifully, and Gets It Wrong - Newgen Resources (DX++)","og_description":"For insurers, explainability is only the beginning. The real test is whether people can still question the machine. There is a question insurance leaders should start asking about AI. Not simply: Can the AI explain its decision? But: What happens when the explanation sounds perfectly reasonable and is still wrong? In an AI-supported underwriting process, […]","og_url":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/","og_site_name":"Newgen Resources (DX++)","article_published_time":"2026-09-25T11:12:19+00:00","article_modified_time":"2026-09-25T11:12:53+00:00","og_image":[{"width":1536,"height":1024,"url":"https:\/\/newgensoft.com\/resources\/wp-content\/uploads\/sites\/6\/2026\/09\/For-Blog.jpg","type":"image\/jpeg"}],"author":"Darshak Parmar","twitter_card":"summary_large_image","twitter_misc":{"Written by":"Darshak Parmar","Est. reading time":"6 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/#article","isPartOf":{"@id":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/"},"author":{"name":"Darshak Parmar","@id":"https:\/\/newgensoft.com\/resources\/#\/schema\/person\/03d87995d5366558c449b58979bdd031"},"headline":"What Happens When AI Explains Itself Beautifully, and Gets It Wrong","datePublished":"2026-09-25T11:12:19+00:00","dateModified":"2026-09-25T11:12:53+00:00","mainEntityOfPage":{"@id":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/"},"wordCount":1257,"commentCount":0,"publisher":{"@id":"https:\/\/newgensoft.com\/resources\/#organization"},"image":{"@id":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/#primaryimage"},"thumbnailUrl":"https:\/\/newgensoft.com\/resources\/wp-content\/uploads\/sites\/6\/2026\/09\/For-Blog.jpg","articleSection":["Blog"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/","url":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/","name":"What Happens When AI Explains Itself Beautifully, and Gets It Wrong - Newgen Resources (DX++)","isPartOf":{"@id":"https:\/\/newgensoft.com\/resources\/#website"},"primaryImageOfPage":{"@id":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/#primaryimage"},"image":{"@id":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/#primaryimage"},"thumbnailUrl":"https:\/\/newgensoft.com\/resources\/wp-content\/uploads\/sites\/6\/2026\/09\/For-Blog.jpg","datePublished":"2026-09-25T11:12:19+00:00","dateModified":"2026-09-25T11:12:53+00:00","breadcrumb":{"@id":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/#primaryimage","url":"https:\/\/newgensoft.com\/resources\/wp-content\/uploads\/sites\/6\/2026\/09\/For-Blog.jpg","contentUrl":"https:\/\/newgensoft.com\/resources\/wp-content\/uploads\/sites\/6\/2026\/09\/For-Blog.jpg","width":1536,"height":1024},{"@type":"BreadcrumbList","@id":"https:\/\/newgensoft.com\/resources\/blog\/what-happens-when-ai-explains-itself-beautifully-and-gets-it-wrong\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/newgensoft.com\/resources\/"},{"@type":"ListItem","position":2,"name":"What Happens When AI Explains Itself Beautifully, and Gets It Wrong"}]},{"@type":"WebSite","@id":"https:\/\/newgensoft.com\/resources\/#website","url":"https:\/\/newgensoft.com\/resources\/","name":"Newgen DX++","description":"","publisher":{"@id":"https:\/\/newgensoft.com\/resources\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/newgensoft.com\/resources\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/newgensoft.com\/resources\/#organization","name":"Newgen Software","url":"https:\/\/newgensoft.com\/resources\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/newgensoft.com\/resources\/#\/schema\/logo\/image\/","url":"https:\/\/newgensoft.com\/resources\/wp-content\/uploads\/sites\/6\/2026\/06\/Newgen-Logo_Light-scaled.png","contentUrl":"https:\/\/newgensoft.com\/resources\/wp-content\/uploads\/sites\/6\/2026\/06\/Newgen-Logo_Light-scaled.png","width":2560,"height":788,"caption":"Newgen Software"},"image":{"@id":"https:\/\/newgensoft.com\/resources\/#\/schema\/logo\/image\/"}},{"@type":"Person","@id":"https:\/\/newgensoft.com\/resources\/#\/schema\/person\/03d87995d5366558c449b58979bdd031","name":"Darshak Parmar","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/ded91f8c09d9e3c95bc58254ae8fa31f63e4d843b2bbe2bbbe15d5254bd1778b?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/ded91f8c09d9e3c95bc58254ae8fa31f63e4d843b2bbe2bbbe15d5254bd1778b?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/ded91f8c09d9e3c95bc58254ae8fa31f63e4d843b2bbe2bbbe15d5254bd1778b?s=96&d=mm&r=g","caption":"Darshak Parmar"}}]}},"_links":{"self":[{"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/posts\/2143532196","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/users\/6"}],"replies":[{"embeddable":true,"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/comments?post=2143532196"}],"version-history":[{"count":3,"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/posts\/2143532196\/revisions"}],"predecessor-version":[{"id":2143532209,"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/posts\/2143532196\/revisions\/2143532209"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/media\/2143532206"}],"wp:attachment":[{"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/media?parent=2143532196"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/categories?post=2143532196"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/tags?post=2143532196"},{"taxonomy":"resources-topic","embeddable":true,"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/resources-topic?post=2143532196"},{"taxonomy":"resources-group","embeddable":true,"href":"https:\/\/newgensoft.com\/resources\/wp-json\/wp\/v2\/resources-group?post=2143532196"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}