<html><head></head><body><div class="ydpebc6f944yahoo-style-wrap" style="font-family:Helvetica Neue, Helvetica, Arial, sans-serif;font-size:13px;"><div></div>
<div dir="ltr" data-setdir="false">The paradigm of an "<span>AI model they’d trained themselves that looks for telltale signals of ChatGPT output" has been proven time and time again to spew out false-positives. College professors have lost their jobs for making unsubstantiated accusations and penalizing students for allegedly using AI or ChatGPT when in fact the students came back with notes and drafts proving definitively that their work was original and not generated by AI. An "accuracy rate" which allows for up to 40% false-positives does not encourage placing any faith in AI models which claim to be able to detect AI produced content. The Swiss team was either lying, or else they were deceiving themselves. Put succinctly, AI can never definitively detect AI.</span></div><div dir="ltr" data-setdir="false"><span><br></span></div><div dir="ltr" data-setdir="false"><span>Nevertheless, there is a very real possibility that at least some LLMs are reusing data sets which were supplied with the help of AI models. The models might very well be feeding upon their own errors. And this does pose a real threat to the integrity of the information presented as output by all LLMs. How LLMs and other AI are trained does influence their behaviors when they are put into use.</span></div><div dir="ltr" data-setdir="false"><span><br></span></div><div dir="ltr" data-setdir="false"><span>Outsourcing any part of AI training does run the risk that the gig workers involved will take every shortcut they can to generate the illusion that they are producing more finished work in a shorter time interval. Gig workers are notorious for doing everything they can to avoid doing any real work.</span></div><div dir="ltr" data-setdir="false"><span><br></span></div><div dir="ltr" data-setdir="false"><span>The obvious solution is this: Pay qualified people to do the training work under the watchful eyes of in-house research staff. It's a lot more expensive, but this is the only way to make sure no one is taking thee easy way out and endangering the integrity of the entire project. When you take shortcuts in technology development, it is your end-users who invariably pay the price. </span></div><div dir="ltr" data-setdir="false"><span><br></span></div><div dir="ltr" data-setdir="false"><span>-- Bob Primak </span></div><div dir="ltr" data-setdir="false"><span><br></span></div><div dir="ltr" data-setdir="false"><span><br></span></div><div dir="ltr" data-setdir="false"><span><br></span></div><div dir="ltr" data-setdir="false"><span><br></span></div><div dir="ltr" data-setdir="false"><span><br></span></div><div dir="ltr" data-setdir="false"><span><br></span></div><div><br></div>
</div><div id="yahoo_quoted_7828991126" class="yahoo_quoted">
<div style="font-family:'Helvetica Neue', Helvetica, Arial, sans-serif;font-size:13px;color:#26282a;">
<div>
On Saturday, June 24, 2023 at 11:43:34 AM EDT, jjrudy1@comcast.net <jjrudy1@comcast.net> wrote:
</div>
<div><br></div>
<div><br></div>
<div><div id="yiv3946723560"><style><!--
#yiv3946723560 filtered {}
#yiv3946723560 filtered {}
#yiv3946723560 filtered {}
#yiv3946723560 filtered {}
#yiv3946723560 p.yiv3946723560MsoNormal, #yiv3946723560 li.yiv3946723560MsoNormal, #yiv3946723560 div.yiv3946723560MsoNormal
{margin:0in;font-size:11.0pt;font-family:"Calibri", sans-serif;}
#yiv3946723560 h1
{margin-right:0in;margin-left:0in;font-size:24.0pt;font-family:"Calibri", sans-serif;font-weight:bold;}
#yiv3946723560 h2
{margin-right:0in;margin-left:0in;font-size:18.0pt;font-family:"Calibri", sans-serif;font-weight:bold;}
#yiv3946723560 h3
{margin-right:0in;margin-left:0in;font-size:13.5pt;font-family:"Calibri", sans-serif;font-weight:bold;}
#yiv3946723560 a:link, #yiv3946723560 span.yiv3946723560MsoHyperlink
{color:#0563C1;text-decoration:underline;}
#yiv3946723560 span.yiv3946723560EmailStyle17
{font-family:"Arial", sans-serif;font-variant:normal !important;color:windowtext;text-transform:none;text-decoration:none none;vertical-align:baseline;}
#yiv3946723560 span.yiv3946723560Heading1Char
{font-family:"Calibri", sans-serif;font-weight:bold;}
#yiv3946723560 span.yiv3946723560Heading2Char
{font-family:"Calibri", sans-serif;font-weight:bold;}
#yiv3946723560 span.yiv3946723560Heading3Char
{font-family:"Calibri", sans-serif;font-weight:bold;}
#yiv3946723560 span.yiv3946723560bylineby--3i70z
{}
#yiv3946723560 p.yiv3946723560bylineauthor--g26rn, #yiv3946723560 li.yiv3946723560bylineauthor--g26rn, #yiv3946723560 div.yiv3946723560bylineauthor--g26rn
{margin-right:0in;margin-left:0in;font-size:11.0pt;font-family:"Calibri", sans-serif;}
#yiv3946723560 span.yiv3946723560screen-reader-text
{}
#yiv3946723560 span.yiv3946723560imagecredit--1fj0h
{}
#yiv3946723560 span.yiv3946723560slideradclosetext--1dj37
{}
#yiv3946723560 .yiv3946723560MsoChpDefault
{font-family:"Calibri", sans-serif;}
#yiv3946723560 filtered {}
#yiv3946723560 div.yiv3946723560WordSection1
{}
#yiv3946723560 filtered {}
#yiv3946723560 filtered {}
#yiv3946723560 filtered {}
#yiv3946723560 filtered {}
#yiv3946723560 filtered {}
#yiv3946723560 filtered {}
#yiv3946723560 filtered {}
#yiv3946723560 filtered {}
#yiv3946723560 filtered {}
#yiv3946723560 filtered {}
#yiv3946723560 ol
{margin-bottom:0in;}
#yiv3946723560 ul
{margin-bottom:0in;}
--></style><div><div class="yiv3946723560WordSection1"><p style="vertical-align:baseline;" class="yiv3946723560MsoNormal"><span style="font-size:12.0pt;font-family:inherit, serif;color:#111111;text-transform:uppercase;"><a rel="nofollow noopener noreferrer" target="_blank" href="https://www.technologyreview.com/topic/artificial-intelligence/"><b><span style="font-family:Courier New;color:white;border:none windowtext 1.0pt;padding:0in;text-decoration:none;">ARTIFICIALThis is scary INTELLIGENCE</span></b></a></span></p><h1 style="vertical-align:baseline;font-style:inherit;font-variant:inherit;font-stretch:inherit;margin:0px;outline:0px;"><span style="font-family:Arial, sans-serif;color:white;">paid to train AI are outsourcing their work… to AI</span></h1><p style="margin:0in;vertical-align:baseline;font-style:inherit;font-variant:inherit;font-weight:inherit;font-stretch:inherit;line-height:inherit;outline:0px;"><b><span style="font-size:18.0pt;font-family:inherit, serif;color:white;border:none windowtext 1.0pt;padding:0in;">It’s a practice that could introduce further errors into already error-prone models.</span></b><span style="font-size:18.0pt;font-family:inherit, serif;color:white;"></span></p><p style="vertical-align:baseline;" class="yiv3946723560MsoNormal"><span class="yiv3946723560bylineby--3i70z"><b><span style="font-size:15.0pt;font-family:inherit, serif;color:white;border:none windowtext 1.0pt;padding:0in;">By </span></b></span><b><span style="font-size:12.0pt;font-family:Arial, sans-serif;color:white;"></span></b></p><ul style="margin-top:0in;" type="disc"><li style="color:white;margin-top:0in;margin-bottom:0in;vertical-align:baseline;font-style:inherit;font-variant:inherit;font-weight:inherit;font-stretch:inherit;line-height:inherit;outline:0px;display:inline-block;" class="yiv3946723560bylineauthor--g26rn"><b><span style="font-size:15.0pt;font-family:inherit, serif;"><a rel="nofollow noopener noreferrer" target="_blank" href="https://www.technologyreview.com/author/rhiannon-williams/"><span style="border:none windowtext 1.0pt;padding:0in;text-decoration:none;">Rhiannon Williams</span><span class="yiv3946723560screen-reader-text"><span style="color:blue;border:none windowtext 1.0pt;padding:0in;text-decoration:none;">archive page</span></span></a></span></b></li></ul><p style="vertical-align:baseline;" class="yiv3946723560MsoNormal"><span style="font-size:12.0pt;font-family:Arial, sans-serif;color:white;">June 22, 2023</span></p><p style="vertical-align:baseline;" class="yiv3946723560MsoNormal"><span style="font-size:12.0pt;font-family:inherit, serif;color:#111111;border:none windowtext 1.0pt;padding:0in;"><img border="0" width="2112" height="1189" style="width:22.0in;min-height:12.3802in;" id="yiv3946723560Picture_x0020_3" src="cid:Vqf8OfFIWa4nXH9z7UMS" yahoo_partid="1.2" alt="anonymous workers at a conveyor belt and strings of words flow out one end"></span><span style="font-size:12.0pt;font-family:inherit, serif;color:#111111;border:none windowtext 1.0pt;padding:0in;"></span></p><p style="vertical-align:baseline;" class="yiv3946723560MsoNormal"><span class="yiv3946723560imagecredit--1fj0h"><span style="font-size:12.0pt;font-family:inherit, serif;color:#616568;text-transform:uppercase;letter-spacing:.75pt;border:none windowtext 1.0pt;padding:0in;">STEPHANIE ARNETT/MITTR | GETTY</span></span><span style="font-size:12.0pt;font-family:inherit, serif;color:#111111;"></span></p><p style="vertical-align:baseline;font-style:inherit;font-variant:inherit;font-stretch:inherit;outline:0px;"><span style="font-size:12.0pt;font-family:Times New Roman, serif;color:black;">A significant proportion of people paid to train AI models may be themselves outsourcing that work to AI, a new study has found. </span></p><p style="vertical-align:baseline;font-style:inherit;font-variant:inherit;font-stretch:inherit;outline:0px;"><span style="font-size:12.0pt;font-family:Times New Roman, serif;color:black;">It takes an incredible amount of data to train AI systems to perform specific tasks accurately and reliably. Many companies pay gig workers on platforms like <a rel="nofollow noopener noreferrer" target="_blank" href="https://www.technologyreview.com/2020/12/11/1014081/ai-machine-learning-crowd-gig-worker-problem-amazon-mechanical-turk/"><span style="border:none windowtext 1.0pt;padding:0in;">Mechanical Turk</span></a> to complete tasks that are typically hard to automate, such as solving CAPTCHAs, labeling data and annotating text. This data is then fed into AI models to train them. The workers are poorly paid and are often expected to complete lots of tasks very quickly. </span></p><div style="border:none;border-top:solid black 1.0pt;padding:0in 0in 0in 0in;"><h2 style="vertical-align:baseline;border:none;padding:0in;border-right-style:initial;border-bottom-style:initial;border-left-style:initial;border-right-color:initial;border-bottom-color:initial;border-left-color:initial;font-style:inherit;font-variant:inherit;font-stretch:inherit;line-height:inherit;padding:0px 0px;outline:0px;"><span style="font-family:inherit, serif;color:#111111;">Related Story</span></h2></div><p style="vertical-align:baseline;" class="yiv3946723560MsoNormal"><a rel="nofollow noopener noreferrer" target="_blank" href="https://www.technologyreview.com/2023/04/04/1070938/we-are-hurtling-toward-a-glitchy-spammy-scammy-ai-powered-internet/"><span style="font-size:10.5pt;font-family:inherit, serif;color:blue;border:none windowtext 1.0pt;padding:0in;text-decoration:none;"><img border="0" width="2112" height="1189" style="width:22.0in;min-height:12.3802in;" id="yiv3946723560Picture_x0020_2" src="cid:5aW0omu0H7MBAAXnOjAQ" yahoo_partid="1.3" alt=""""></span></a><span style="font-size:10.5pt;font-family:inherit, serif;color:#111111;"></span></p><p style="vertical-align:baseline;" class="yiv3946723560MsoNormal"><span style="font-size:10.5pt;font-family:inherit, serif;color:#111111;"><a rel="nofollow noopener noreferrer" target="_blank" href="https://www.technologyreview.com/2023/04/04/1070938/we-are-hurtling-toward-a-glitchy-spammy-scammy-ai-powered-internet/"><b><span style="font-size:11.0pt;font-family:Calibri, sans-serif;border:none windowtext 1.0pt;padding:0in;"></span></b></a></span></p><p style="vertical-align:baseline;" class="yiv3946723560MsoNormal"><b><u><span style="font-size:10.5pt;font-family:inherit, serif;color:blue;border:none windowtext 1.0pt;padding:0in;"><a rel="nofollow noopener noreferrer" target="_blank" href="https://www.technologyreview.com/2023/04/04/1070938/we-are-hurtling-toward-a-glitchy-spammy-scammy-ai-powered-internet/"><span style="color:blue;">We are hurtling toward a glitchy, spammy, scammy, AI-powered internet</span><span style="font-size:11.0pt;font-family:Calibri, sans-serif;color:windowtext;border:none;font-weight:normal;text-decoration:none;"></span></a></span></u></b></p><p style="vertical-align:baseline;" class="yiv3946723560MsoNormal"><span style="font-size:10.5pt;font-family:inherit, serif;color:#111111;"> </span></p><p style="margin:0in;vertical-align:baseline;font-style:inherit;font-variant:inherit;font-weight:inherit;font-stretch:inherit;line-height:inherit;outline:0px;"><span style="font-size:10.5pt;font-family:inherit, serif;color:#111111;">Large language models are full of security vulnerabilities, yet they’re being embedded into tech products on a vast scale.</span></p><p style="vertical-align:baseline;font-style:inherit;font-variant:inherit;font-stretch:inherit;outline:0px;"><span style="font-size:12.0pt;font-family:Times New Roman, serif;color:black;">No wonder some of them may be turning to tools like ChatGPT to maximize their earning potential. But how many? To find out, a team of researchers from the Swiss Federal Institute of Technology (EPFL) hired 44 people on the gig work platform Amazon Mechanical Turk to summarize 16 extracts from medical research papers. Then they analyzed their responses using an AI model they’d trained themselves that looks for telltale signals of ChatGPT output, such as lack of variety in choice of words. They also extracted the workers’ keystrokes in a bid to work out whether they’d copied and pasted their answers, an indicator that they’d generated their responses elsewhere.</span></p><p style="vertical-align:baseline;font-style:inherit;font-variant:inherit;font-stretch:inherit;outline:0px;"><span style="font-size:12.0pt;font-family:Times New Roman, serif;color:black;">They estimated that somewhere between 33% and 46% of the workers had used AI models like OpenAI’s ChatGPT. It’s a percentage that’s likely to grow even higher as ChatGPT and other AI systems become more powerful and easily accessible, according to the authors of the study, which has been shared on <a rel="nofollow noopener noreferrer" target="_blank" href="https://arxiv.org/pdf/2306.07899.pdf"><span style="border:none windowtext 1.0pt;padding:0in;">arXiv</span></a> and is yet to be peer-reviewed. </span></p><p style="vertical-align:baseline;font-style:inherit;font-variant:inherit;font-stretch:inherit;outline:0px;"><span style="font-size:12.0pt;font-family:Times New Roman, serif;color:black;">“I don’t think it’s the end of crowdsourcing platforms. It just changes the dynamics,” says Robert West, an assistant professor at EPFL, who coauthored the study. </span></p><p style="vertical-align:baseline;font-style:inherit;font-variant:inherit;font-stretch:inherit;outline:0px;"><span style="font-size:12.0pt;font-family:Times New Roman, serif;color:black;">Using AI-generated data to train AI could introduce further errors into already error-prone models. Large language models regularly present false information as fact. If they generate incorrect output that is itself used to train other AI models, the errors can be absorbed by those models and amplified over time, making it more and more difficult to work out their origins, says Ilia Shumailov, a junior research fellow in computer science at Oxford University, who was not involved in the project.</span></p><p style="vertical-align:baseline;font-style:inherit;font-variant:inherit;font-stretch:inherit;outline:0px;"><span style="font-size:12.0pt;font-family:Times New Roman, serif;color:black;">Even worse, there’s no simple fix. “The problem is, when you’re using artificial data, you acquire the errors from the misunderstandings of the models and statistical errors,” he says. “You need to make sure that your errors are not biasing the output of other models, and there’s no simple way to do that.”</span></p><p style="vertical-align:baseline;font-style:inherit;font-variant:inherit;font-stretch:inherit;outline:0px;"><span style="font-size:12.0pt;font-family:Times New Roman, serif;color:black;">The study highlights the need for new ways to check whether data has been produced by humans or AI. It also highlights one of the problems with tech companies’ tendency to rely on gig workers to do the vital work of tidying up the data fed to AI systems. </span></p><p style="vertical-align:baseline;font-style:inherit;font-variant:inherit;font-stretch:inherit;outline:0px;"><span style="font-size:12.0pt;font-family:Times New Roman, serif;color:black;">“I don’t think everything will collapse,” says West. “But I think the AI community will have to investigate closely which tasks are most prone to being automated and to work on ways to prevent this.”</span></p><p style="vertical-align:baseline;" class="yiv3946723560MsoNormal"><span class="yiv3946723560slideradclosetext--1dj37"><span style="font-size:12.0pt;font-family:inherit, serif;color:#111111;border:none windowtext 1.0pt;padding:0in;">hide</span></span><span style="font-size:12.0pt;font-family:inherit, serif;color:#111111;"></span></p><h3 style="margin:0in;vertical-align:baseline;font-style:inherit;font-variant:inherit;font-stretch:inherit;line-height:inherit;outline:0px;"><span style="font-size:18.0pt;font-family:inherit, serif;color:#111111;">by <a rel="nofollow noopener noreferrer" target="_blank" href="https://www.technologyreview.com/author/rhiannon-williams/"><span style="border:none windowtext 1.0pt;padding:0in;text-decoration:none;">Rhiannon Williams</span></a></span></h3><p class="yiv3946723560MsoNormal"><span style="font-size:12.0pt;font-family:Arial, sans-serif;"> </span></p><p class="yiv3946723560MsoNormal"><span style="font-size:12.0pt;font-family:Arial, sans-serif;"> </span></p><p class="yiv3946723560MsoNormal"><span style="font-size:10.0pt;font-family:Arial, sans-serif;">John Rudy</span></p><p class="yiv3946723560MsoNormal"><span style="font-size:10.0pt;font-family:Arial, sans-serif;"> </span></p><p class="yiv3946723560MsoNormal"><span style="font-size:10.0pt;font-family:Arial, sans-serif;">781-861-0402</span></p><p class="yiv3946723560MsoNormal"><span style="font-size:10.0pt;font-family:Arial, sans-serif;">781-718-8334 cell</span></p><p class="yiv3946723560MsoNormal"><span style="font-size:10.0pt;font-family:Arial, sans-serif;">13 Hawthorne Lane</span></p><p class="yiv3946723560MsoNormal"><span style="font-size:10.0pt;font-family:Arial, sans-serif;">Bedford MA</span></p><p class="yiv3946723560MsoNormal"><span style="font-size:10.0pt;font-family:Arial, sans-serif;"><a rel="nofollow noopener noreferrer" ymailto="mailto:jjrudy1@comcast.net" target="_blank" href="mailto:jjrudy1@comcast.net">jjrudy1@comcast.net</a></span></p><p class="yiv3946723560MsoNormal"><img border="0" width="124" height="115" style="width:1.2916in;min-height:1.1927in;" id="yiv3946723560Picture_x0020_1" src="cid:w1MTgGmzoW0FsWL1wJ4P" yahoo_partid="1.4"><span style="font-size:10.0pt;font-family:Arial, sans-serif;"></span></p><p class="yiv3946723560MsoNormal"> </p></div></div></div>===============================================<br>::The Lexington Computer and Technology Group Mailing List::<br>Reply goes to sender only; Reply All to send to list.<br>Send to the list: <a ymailto="mailto:LCTG@lists.toku.us" href="mailto:LCTG@lists.toku.us">LCTG@lists.toku.us</a> Message archives: <a href="http://lists.toku.us/pipermail/lctg-toku.us/" target="_blank">http://lists.toku.us/pipermail/lctg-toku.us/</a><br>To subscribe: email <a ymailto="mailto:lctg-subscribe@toku.us" href="mailto:lctg-subscribe@toku.us">lctg-subscribe@toku.us</a> To unsubscribe: email <a ymailto="mailto:lctg-unsubscribe@toku.us" href="mailto:lctg-unsubscribe@toku.us">lctg-unsubscribe@toku.us</a><br>Future and Past meeting information: <a href="http://LCTG.toku.us" target="_blank">http://LCTG.toku.us</a><br>List information: <a href="http://lists.toku.us/listinfo.cgi/lctg-toku.us" target="_blank">http://lists.toku.us/listinfo.cgi/lctg-toku.us</a><br>This message was sent to <a ymailto="mailto:bobprimak@yahoo.com." href="mailto:bobprimak@yahoo.com.">bobprimak@yahoo.com.</a><br>Set your list options: <a href="http://lists.toku.us/options.cgi/lctg-toku.us/bobprimak@yahoo.com" target="_blank">http://lists.toku.us/options.cgi/lctg-toku.us/bobprimak@yahoo.com</a><br></div>
</div>
</div></body></html>