Daily Archives: January 5, 2023
The Expanding Dark Forest and Generative AI
Proving you’re a human on a web flooded with generative AI content
The dark forest theory of the web points to the increasingly life-like but life-less state of being online.
Most open and publicly available spaces on the web are overrun with bots, advertisers, trolls, data scrapers, clickbait, keyword-stuffing “content creators,” and algorithmically manipulated junk.
It’s like a dark forest that seems eerily devoid of human life – all the living creatures are hidden beneath the ground or up in trees. If they reveal themselves, they risk being attacked by automated predators. Read More
Achieving Individual — and Organizational — Value With AI
Findings from the 2022 Artificial Intelligence and Business Strategy Global Executive Study and Research Project
New research shows that employees derive individual value from AI when using the technology improves their sense of competency, autonomy, and relatedness. Likewise, organizations are far more likely to obtain value from AI when their workers do. This report offers key insights for leaders on achieving individual and organizational value with artificial intelligence in their organizations. Read More
AI legal assistant will help defendant fight a speeding case in court
In February, an AI from DoNotPay is set to tell a defendant exactly what to say and when during an entire court case. It is likely to be the first ever case defended by an artificial intelligence
An artificial intelligence is set to advise a defendant in court for the first time ever. The AI will run on a smartphone and listen to all speech in the courtroom in February before instructing the defendant on what to say via an earpiece.
The location of the court and the name of the defendant are being kept under wraps by DoNotPay, the company that created the AI. But it is understood that the defendant is charged with speeding and that they will say only what DoNotPay’s tool tells them to via an earbud. The case is being considered as a test by the company, which has agreed to pay any fines, should they be imposed, says the firm’s founder, Joshua Browder. Read More
An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion
Text-to-image models offer unprecedented freedom to guide creation through natural language. Yet, it is unclear how such freedom can be exercised to generate images of specific unique concepts, modify their appearance, or compose them in new roles and novel scenes. In other words, we ask: how can we use language-guided models to turn our cat into a painting, or imagine a new product based on our favorite toy? Here we present a simple approach that allows such creative freedom. Using only 3-5 images of a user-provided concept, like an object or a style, we learn to represent it through new “words” in the embedding space of a frozen text-to-image model. These “words” can be composed into natural language sentences, guiding personalized creation in an intuitive way. Notably, we find evidence that a single word embedding is sufficient for capturing unique and varied concepts. We compare our approach to a wide range of baselines, and demonstrate that it can more faithfully portray the concepts across a range of applications and tasks. Read More
Deepfake Text Detector Tool GPTZero Spots AI Writing
A new tool is attempting to spot when text is written by ChatGPT and other generative AI engines. Princeton student and former open source investigator for BBC Africa Eye Edward Tian created GPTZero to identify deepfake text, a subject attracting a growing amount of interest in the academic and business world as the debate over how to respond to the potential misuse of AI continues.
Tian’s app processes submitted text for indicators of AI origins like randomness and complexity in how it is written, technically referred to as “perplexity and burstiness.” GPTZero was popular enough to almost immediately crash the hosting website, but you can play with it online here. … Voicebot ran multiple tests of GPTZero using six different generative AI tools, including ChatGPT, a few GPT-3 derived tools, and AI21. Tian’s creation caught the AI-generated text every time and correctly identified text written by a human in more than a dozen cases. Tian doesn’t have enough data to measure accuracy yet, though he said he is working on publishing one. Not bad for an app thrown together on New Year’s Eve. Read More
Will we run out of data? An analysis of the limits of scaling datasets in Machine Learning
We analyze the growth of dataset sizes used in machine learning for natural language processing and computer vision, and extrapolate these using two methods: using the historical growth rate and estimating the compute-optimal dataset size for future predicted compute budgets. We investigate the growth in data usage by estimating the total stock of unlabeled data available on the internet over the coming decades. Our analysis indicates that the stock of high-quality language data will be exhausted soon; likely before 2026. By contrast, the stock of low-quality language data and image data will be exhausted only much later; between 2030 and 2050 (for low-quality language) and between 2030 and 2060 (for images). Our work suggests that the current trend of ever-growing ML models that rely on enormous datasets might slow down if data efficiency is not drastically improved or new sources of data become available. Read More
#machine-learning
Rick's Cafe AI 