{"id":2206,"date":"2025-11-19T16:42:26","date_gmt":"2025-11-19T16:42:26","guid":{"rendered":"https:\/\/retrotradingreport.com\/index.php\/2025\/11\/19\/why-overconfident-ai-models-are-prioritising-assertiveness-over-accuracy\/"},"modified":"2025-11-19T16:42:26","modified_gmt":"2025-11-19T16:42:26","slug":"why-overconfident-ai-models-are-prioritising-assertiveness-over-accuracy","status":"publish","type":"post","link":"https:\/\/retrotradingreport.com\/index.php\/2025\/11\/19\/why-overconfident-ai-models-are-prioritising-assertiveness-over-accuracy\/","title":{"rendered":"Why overconfident AI models are prioritising assertiveness over accuracy"},"content":{"rendered":"
<\/div>\n

Concerns are mounting over the reliability of artificial intelligence models, as new research reveals that some popular systems produce incorrect information in over a third of their responses, despite the increasing reliance and trust placed in AI technology, ING Group said on Wednesday.<\/p>\n

Modern AI models, featuring deep reasoning, long-term memory, and autonomous agents, can perform tasks like web browsing with minimal human intervention. <\/p>\n

However, the execution of these tasks demands extensive data, leading to a greater dependence on external data sources that are often uncontrolled and unverified, ING\u2019s Julian Geib, junior economist, global trade, said in a report.<\/p>\n

Overconfidence<\/h2>\n
\n

This increased exposure can lead to behaviour that resembles overconfidence, a cognitive bias where confidence in one\u2019s knowledge or abilities exceeds actual accuracy. <\/p>\n<\/blockquote>\n

Leading AI systems generate false claims at a rate of up to 40%, a consequence highlighted in a recent study by the European Broadcasting Union (EBU).<\/p>\n

The increased frequency of responses aligns with a shift in AI model behavior. <\/p>\n

Earlier AI systems were programmed to refuse to answer queries regarding topics outside their training datasets. <\/p>\n

However, contemporary systems with web connectivity are engineered to answer more often, even when the information available is limited or uncertain.<\/p>\n

Increased user engagement is a benefit, but it results in more fabricated output, which we term “AI hallucinations”, Geib said. <\/p>\n

\n

However, the models often deliver such responses with strong confidence, creating the impression that they are unquestionably correct.<\/p>\n

Source: ING Research<\/figcaption><\/figure>\n<\/blockquote>\n

Fluency over accuracy<\/h2>\n

Even newer AI models frequently experience hallucinations for several reasons. <\/p>\n

Primarily, when users pose vague or overly complex questions, the model struggles with interpretation. <\/p>\n

This often leads the model to rely on statistical patterns to “fill in the blanks,” generating a seemingly complete, but potentially factually inaccurate, response, Geib said. <\/p>\n

Although these responses aim to be helpful, they can introduce incorrect information.<\/p>\n

Fine-tuning models with human feedback often favours confident, helpful-sounding answers, leading to a bias toward inaccurate but assertive statements over cautious or uncertain responses.<\/p>\n

The problem is worsened by the plummeting “no response rate.” <\/p>\n

Older models refused nearly 40% of queries, but newer ones answer almost everything. <\/p>\n

In critical fields like politics and health, this prioritisation of fluency over accuracy creates serious misinformation risks.<\/p>\n

AI is becoming a more common tool for accessing information on current events, especially among younger demographics. <\/p>\n

Notably, 15% of people under the age of 25 state that they rely on AI chatbots as their main source for news.<\/p>\n

\u201cGiven the rising usage of AI both privately and in businesses, accuracy should be a priority,\u201d Geib said. <\/p>\n

Awareness vital<\/h2>\n

Geib added:<\/p>\n

\n

\u201cIn Germany, there’s a saying: “stiffly claimed is already half proven” \u2013 but confidence does not automatically translate into correctness.<\/p>\n<\/blockquote>\n

The current limitations in the accuracy of AI suggest that the wholesale replacement of entire professional fields in the immediate future is a significantly low probability event, according to Geib. <\/p>\n

This is primarily because human professionals, in most domains, operate with a degree of nuanced judgment, contextual understanding, and accuracy that current AI systems struggle to consistently replicate. <\/p>\n

The risk of widespread job displacement, therefore, becomes critical only in scenarios where the practitioners within a profession become entirely reliant upon\u2014and fail to critically verify\u2014the potentially flawed or erroneous data and conclusions generated by AI. <\/p>\n

Essentially, AI currently serves as a powerful, yet imperfect, tool, and its inaccuracies ensure that human oversight, critical thinking, and validation remain indispensable components of professional work.<\/p>\n

\u201cAI-generated statements should be treated with the same critical mindset as human claims,\u201d Geib noted. <\/p>\n

The post Why overconfident AI models are prioritising assertiveness over accuracy<\/a> appeared first on Invezz<\/a><\/p>\n

<\/p>\n","protected":false},"excerpt":{"rendered":"

Concerns are mounting over the reliability of artificial intelligence models, as new research reveals that some popular systems produce incorrect information in over a third of their responses, despite the increasing reliance and trust placed in AI technology, ING Group said on Wednesday.Modern AI models, featuring deep reasoning, long-term memory, and autonomous agents, can perform…<\/p>\n","protected":false},"author":1,"featured_media":2207,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3],"tags":[],"class_list":["post-2206","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-investing"],"_links":{"self":[{"href":"https:\/\/retrotradingreport.com\/index.php\/wp-json\/wp\/v2\/posts\/2206","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/retrotradingreport.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/retrotradingreport.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/retrotradingreport.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/retrotradingreport.com\/index.php\/wp-json\/wp\/v2\/comments?post=2206"}],"version-history":[{"count":0,"href":"https:\/\/retrotradingreport.com\/index.php\/wp-json\/wp\/v2\/posts\/2206\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/retrotradingreport.com\/index.php\/wp-json\/wp\/v2\/media\/2207"}],"wp:attachment":[{"href":"https:\/\/retrotradingreport.com\/index.php\/wp-json\/wp\/v2\/media?parent=2206"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/retrotradingreport.com\/index.php\/wp-json\/wp\/v2\/categories?post=2206"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/retrotradingreport.com\/index.php\/wp-json\/wp\/v2\/tags?post=2206"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}