New Review Essay on @lmesseri tremendous new book, ethnography & tech, social hopes, & false dreams of tech solutionism. Also discussing work of Andrew Brock, Zeynep Tufekci & Kelsie Nabben on Black Twitter, Twitter & ethnographies of DAOs.
Backstabbing, bluffing and playing dead: has AI learned to deceive? – podcast
“Dr Peter Park, an AI existential safety researcher at MIT and author of the research, tells Ian Sample about the different examples of deception he uncovered, and why they will be so difficult to tackle as long as AI remains a black box.”
"A Bayesian analysis showed that participants had high expectations and performed descriptively better irrespective of the AI description when a sham-AI was present. Using cognitive modeling, we could trace this advantage back to participants gathering more information."
Agnes Mercedes Kloft, Robin Welsch, Thomas Kosch, and Steeven Villa. 2024. "AI enhances our performance, I have no doubt this one will do the same": The Placebo effect is robust to negative descriptions of AI. In Proceedings of the CHI Conference on Human Factors in Computing Systems (CHI '24). Association for Computing Machinery, New York, NY, USA, Article 299, 1–24. https://doi.org/10.1145/3613904.3642633
"A Bayesian analysis showed that participants had high expectations and performed descriptively better irrespective of the AI description when a sham-AI was present. Using cognitive modeling, we could trace this advantage back to participants gathering more information."
Agnes Mercedes Kloft, Robin Welsch, Thomas Kosch, and Steeven Villa. 2024. "AI enhances our performance, I have no doubt this one will do the same": The Placebo effect is robust to negative descriptions of AI. In Proceedings of the CHI Conference on Human Factors in Computing Systems (CHI '24). Association for Computing Machinery, New York, NY, USA, Article 299, 1–24. https://doi.org/10.1145/3613904.3642633
"A Bayesian analysis showed that participants had high expectations and performed descriptively better irrespective of the AI description when a sham-AI was present. Using cognitive modeling, we could trace this advantage back to participants gathering more information."
Agnes Mercedes Kloft, Robin Welsch, Thomas Kosch, and Steeven Villa. 2024. "AI enhances our performance, I have no doubt this one will do the same": The Placebo effect is robust to negative descriptions of AI. In Proceedings of the CHI Conference on Human Factors in Computing Systems (CHI '24). Association for Computing Machinery, New York, NY, USA, Article 299, 1–24. https://doi.org/10.1145/3613904.3642633
"A Bayesian analysis showed that participants had high expectations and performed descriptively better irrespective of the AI description when a sham-AI was present. Using cognitive modeling, we could trace this advantage back to participants gathering more information."
Agnes Mercedes Kloft, Robin Welsch, Thomas Kosch, and Steeven Villa. 2024. "AI enhances our performance, I have no doubt this one will do the same": The Placebo effect is robust to negative descriptions of AI. In Proceedings of the CHI Conference on Human Factors in Computing Systems (CHI '24). Association for Computing Machinery, New York, NY, USA, Article 299, 1–24. https://doi.org/10.1145/3613904.3642633
AI deception: A survey of examples, risks, and potential solutions
"Large language models and other AI systems have already learned, from their training, the ability to deceive via techniques such as manipulation, sycophancy, and cheating the safety test. AI’s increasing capabilities at deception pose serious risks, ranging from short-term risks, such as fraud and election tampering, to long-term risks, such as losing control of AI systems."
AI deception: A survey of examples, risks, and potential solutions
"Large language models and other AI systems have already learned, from their training, the ability to deceive via techniques such as manipulation, sycophancy, and cheating the safety test. AI’s increasing capabilities at deception pose serious risks, ranging from short-term risks, such as fraud and election tampering, to long-term risks, such as losing control of AI systems."
AI deception: A survey of examples, risks, and potential solutions
“Large language models and other AI systems have already learned, from their training, the ability to deceive via techniques such as manipulation, sycophancy, and cheating the safety test. AI’s increasing capabilities at deception pose serious risks, ranging from short-term risks, such as fraud and election tampering, to long-term risks, such as losing control of AI systems.”
AI deception: A survey of examples, risks, and potential solutions
“Large language models and other AI systems have already learned, from their training, the ability to deceive via techniques such as manipulation, sycophancy, and cheating the safety test. AI’s increasing capabilities at deception pose serious risks, ranging from short-term risks, such as fraud and election tampering, to long-term risks, such as losing control of AI systems.”
AI deception: A survey of examples, risks, and potential solutions
"Large language models and other AI systems have already learned, from their training, the ability to deceive via techniques such as manipulation, sycophancy, and cheating the safety test. AI’s increasing capabilities at deception pose serious risks, ranging from short-term risks, such as fraud and election tampering, to long-term risks, such as losing control of AI systems."
AI deception: A survey of examples, risks, and potential solutions
"Large language models and other AI systems have already learned, from their training, the ability to deceive via techniques such as manipulation, sycophancy, and cheating the safety test. AI’s increasing capabilities at deception pose serious risks, ranging from short-term risks, such as fraud and election tampering, to long-term risks, such as losing control of AI systems."
Congratulations to Harvard University History of Science doctoral candidate Aaron Gluck-Thaler on the 2024-25 CBI Tomash Fellowship. We are thrilled to have Aaron as a fellow in the upcoming academic year! #ai#artificialintelligence#surveillance#science#tech#history
DeepMind’s AI can ‘predict how all of life’s molecules interact with each other’
"AlphaFold 3 is able to envision how the complex shapes and networks of molecules – present in every cell in the human body – are connected and how the smallest of changes in these can affect biological functions that can lead to diseases."
‘Second renaissance’: tech uncovers ancient scroll secrets of Plato and co
"The project belongs to a new wave of efforts that seek to read, restore and translate ancient and even lost languages with cutting-edge technologies. Armed with modern tools, many powered by artificial intelligence, scholars are starting to read what had long been considered unreadable."
"Even though it was hoped that machines might overcome human bias, this assumption often fails due to a problematic or theoretically implausible selection of variables that are fed into the model and because of small size, low representativeness, and presence of bias in the training data [5.]."
Suchotzki, K. and Gamer, M. (2024) 'Detecting deception with artificial intelligence: promises and perils,' Trends in Cognitive Sciences [Preprint]. https://doi.org/10.1016/j.tics.2024.04.002.
"While ChatGPT-4 correlates closely with established risk stratification tools regarding mean scores, its inconsistency when presented with identical patient data on separate occasions raises concerns about its reliability."
Heston TF, Lewis LM (2024) ChatGPT provides inconsistent risk-stratification of patients with atraumatic chest pain. PLOS ONE 19(4): e0301854. https://doi.org/10.1371/journal.pone.0301854