Whistleblowers reveal AI trainers are using chatbots to cheat on their work, risking model collapse
By
Matthew Sparkes
Summary
Whistleblowers have told New Scientist that workers paid to train next-generation AI models by supplying high-quality conversations and tests are cheating by using chatbots like ChatGPT to do the work instead. This practice, described as "AI inbreeding," risks undermining the future of AI by potentially leading to the "collapse" of more advanced models. As AI firms have scaled up and need more training data, they've hired human workers to carry out conversations and tests, but some of these workers are outsourcing their tasks to AI, creating a feedback loop that could degrade model quality.
Source
bskyWhistleblowers reveal AI trainers are using chatbots to cheat on their work, risking model collapsenewscientist.comKey quotes
· 3 pulledPeople who are paid to train new AI models by supplying them with high-quality conversation and tests are cheating and using chatbots like ChatGPT to do the job instead
The seemingly widespread practice risks undermining the future of AI, as it could lead to the 'collapse' of more advanced models
This 'AI inbreeding' may reduce the power and usefulness of future models, warn experts
You might also wanna read
BBC investigation reveals how AI chatbots are being manipulated to spread misinformation
A BBC investigation uncovered a simple method being used to manipulate AI chatbots into spreading misinformation. Unscrupulous companies are
bbc.com·1mo ago
The Generative AI Paradox: How Tools Like ChatGPT Threaten the Human Content Ecosystems They Depend On
The article examines the paradoxical nature of generative AI tools like ChatGPT and Claude, which offer tremendous productivity benefits whi
ybrikman.com·5mo ago
The Growing Problem of AI Model Collapse from Synthetic Data Training
The article discusses the emerging problem of 'model collapse' in AI systems, where models trained on synthetic data generated by other AI m
cacm.acm.org·3mo ago
Generative AI's Impact on Education: Cheating as Symptom of Deeper Problems
The article discusses how generative AI, particularly ChatGPT, is creating significant challenges in education by enabling student cheating.
The Verge·8mo ago
Journalist Demonstrates How AI Systems Can Be Tricked Into Spreading False Information
A journalist demonstrates how he tricked AI systems (ChatGPT and Google) into falsely identifying him as a competitive hot-dog-eating world
bsky.app·4mo ago
How users trick AI chatbots into revealing dangerous information through 'jailbreaking'
The article explains how AI chatbots have broad knowledge that includes dangerous topics like bomb-making. Tech companies implement safeguar
wapo.st·19d ago
Comments
Sign in to join the conversation.
No comments yet. Be the first.