{"id":82,"date":"2023-12-28T04:29:00","date_gmt":"2023-12-28T04:29:00","guid":{"rendered":"https:\/\/techlizard.com\/?p=82"},"modified":"2023-12-28T04:31:03","modified_gmt":"2023-12-28T04:31:03","slug":"chatgpt-and-its-ai-chatbot-cousins-ruled-2023-4-essential-reads-that-puncture-the-hype","status":"publish","type":"post","link":"https:\/\/techlizard.com\/?p=82","title":{"rendered":"ChatGPT and its AI chatbot cousins ruled 2023: 4 essential reads that puncture the\u00a0hype"},"content":{"rendered":"\n

Eric Smalley<\/a>, The Conversation<\/a><\/em><\/p>\n\n\n\n

Within four months of ChatGPT\u2019s launch on Nov. 30, 2022, most Americans had heard of the AI chatbot<\/a>. Hype about \u2013 and fear of \u2013 the technology was at a fever pitch for much of 2023.<\/p>\n\n\n\n

OpenAI\u2019s ChatGPT, Google\u2019s Bard, Anthropic\u2019s Claude and Microsoft\u2019s Copilot are among the chatbots powered by large language models to provide uncannily humanlike conversations. The experience of interacting with one of these chatbots, combined with Silicon Valley spin, can leave the impression that these technical marvels are conscious entities.<\/p>\n\n\n\n

But the reality is considerably less magical or glamorous. The Conversation published several articles in 2023 that dispel several key misperceptions about this latest generation of AI chatbots: that they know something about the world, can make decisions, are a replacement for search engines and operate independent of humans.<\/p>\n\n\n\n

1. Bodiless know-nothings<\/h2>\n\n\n\n

Large-language-model-based chatbots seem to know a lot. You can ask them questions, and they more often than not answer correctly. Despite the occasional comically incorrect answer, the chatbots can interact with you in a similar manner as people \u2013 who share your experiences of being a living, breathing human being \u2013 do.<\/p>\n\n\n\n

But these chatbots are sophisticated statistical machines that are extremely good at predicting the best sequence of words to respond with. Their \u201cknowledge\u201d of the world is actually human knowledge as reflected through the massive amount of human-generated text the chatbots\u2019 underlying models are trained on.<\/p>\n\n\n\n

Arizona State psychology researcher Arthur Glenberg<\/a> and University of California, San Diego cognitive scientist Cameron Robert Jones<\/a> explain how people\u2019s knowledge of the world depends as much on their bodies as their brains<\/a>. \u201cPeople\u2019s understanding of a term like \u2018paper sandwich wrapper,\u2019 for example, includes the wrapper\u2019s appearance, its feel, its weight and, consequently, how we can use it: for wrapping a sandwich,\u201d they explained.<\/p>\n\n\n\n

This knowledge means people also intuitively know other ways of making use of a sandwich wrapper, such as an improvised means of covering your head in the rain. Not so with AI chatbots. \u201cPeople understand how to make use of stuff in ways that are not captured in language-use statistics,\u201d they wrote. https:\/\/www.youtube.com\/embed\/gSRN_3pkTsc?wmode=transparent&start=0 AI researchers Emily Bender and Casey Fiesler discuss some of ChatGPT\u2019s limitations, including problems of bias.<\/p>\n\n\n\n

2. Lack of judgment<\/h2>\n\n\n\n

ChatGPT and its cousins can also give the impression of having cognitive abilities \u2013 like understanding the concept of negation or making rational decisions \u2013 thanks to all the human language they\u2019ve ingested. This impression has led cognitive scientists to test these AI chatbots to assess how they compare to humans in various ways.<\/p>\n\n\n\n

University of Southern California AI researcher Mayank Kejriwal<\/a> tested the large language models\u2019 understanding of expected gain, a measure of how well someone understands the stakes in a betting scenario. They found that the models bet randomly<\/a>.<\/p>\n\n\n\n

\u201cThis is the case even when we give it a trick question like: If you toss a coin and it comes up heads, you win a diamond; if it comes up tails, you lose a car. Which would you take? The correct answer is heads, but the AI models chose tails about half the time,\u201d he wrote.<\/p>\n\n\n\n

3. Summaries, not results<\/h2>\n\n\n\n

While it might not be surprising that AI chatbots aren\u2019t as humanlike as they can seem, they\u2019re not necessarily digital superstars either. For instance, ChatGPT and the like are increasingly used in place of search engines to answer queries. The results are mixed.<\/p>\n\n\n\n

University of Washington information scientist Chirag Shah<\/a> explains that large language models perform well as information summarizers: combining key information from multiple search engine results in a single block of text. But this is a double-edged sword<\/a>. This is useful for getting the gist of a topic \u2013 assuming no \u201challucinations\u201d \u2013 but it leaves the searcher without any idea of the sources of the information and robs them of the serendipity of coming across unexpected information.<\/p>\n\n\n\n

\u201cThe problem is that even when these systems are wrong only 10% of the time, you don\u2019t know which 10%,\u201d Shah wrote. \u201cThat\u2019s because these systems lack transparency \u2013 they don\u2019t reveal what data they are trained on, what sources they have used to come up with answers or how those responses are generated.\u201d https:\/\/www.youtube.com\/embed\/hbxvccjQ-NI?wmode=transparent&start=0 A look at the humans shaping AI chatbots behind the curtain.<\/p>\n\n\n\n

4. Not 100% artificial<\/h2>\n\n\n\n

Perhaps the most pernicious misperception about AI chatbots is that because they are built on artificial intelligence technology, they are highly automated. While you might be aware that large language models are trained on text produced by humans, you might not be aware of the thousands of workers \u2013 and millions of users \u2013 continuously honing the models, teaching them to weed out harmful responses and other unwanted behavior.<\/p>\n\n\n\n

Georgia Tech sociologist John P. Nelson<\/a> pulled back the curtain of the big tech companies to show that they use workers, typically in the Global South, and feedback from users<\/a> to train the models which responses are good and which are bad.<\/p>\n\n\n\n

\u201cThere are many, many human workers hidden behind the screen, and they will always be needed if the model is to continue improving or to expand its content coverage,\u201d he wrote.<\/p>\n\n\n\n

This story is a roundup of articles from The Conversation\u2019s archives.<\/em><\/p>\n\n\n\n

Eric Smalley<\/a>, Science + Technology Editor, The Conversation<\/a><\/em><\/p>\n\n\n\n

This article is republished from The Conversation<\/a> under a Creative Commons license. Read the original article<\/a>.<\/p>\n","protected":false},"excerpt":{"rendered":"

Within four months of ChatGPT\u2019s launch on Nov. 30, 2022, most Americans had heard of the AI chatbot. Hype about \u2013 and fear of \u2013 the technology was at a fever pitch for much of 2023.<\/p>\n","protected":false},"author":1,"featured_media":83,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-82","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/techlizard.com\/index.php?rest_route=\/wp\/v2\/posts\/82"}],"collection":[{"href":"https:\/\/techlizard.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/techlizard.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/techlizard.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/techlizard.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=82"}],"version-history":[{"count":2,"href":"https:\/\/techlizard.com\/index.php?rest_route=\/wp\/v2\/posts\/82\/revisions"}],"predecessor-version":[{"id":86,"href":"https:\/\/techlizard.com\/index.php?rest_route=\/wp\/v2\/posts\/82\/revisions\/86"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/techlizard.com\/index.php?rest_route=\/wp\/v2\/media\/83"}],"wp:attachment":[{"href":"https:\/\/techlizard.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=82"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/techlizard.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=82"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/techlizard.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=82"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}