News of the Cyber World

[kyonides]

AI Fails Pediatric Diagnosis Test Miserably with 17% Accuracy Rating

Ars Technica reports ChatGPT-4, OpenAI’s most advanced AI model, recently faced a major test at the hands of researchers at Cohen Children’s Medical Center in New York. The model was tested against 100 pediatric case challenges, sourced from prominent medical journals JAMA Pediatrics and NEJM, covering the period between 2013 and 2023. These cases are known for their complexity, serving as educational challenges for practicing physicians to diagnose conditions based on the given information.

The AI’s performance was extremely poor on the test, achiving an accurate rate of just 17 percent. Out of the 100 cases, ChatGPT-4 correctly diagnosed only 17. It made incorrect diagnoses in 72 cases, and in 11 cases, it provided answers that were too broad or unspecific to be considered correct. For instance, in one case, ChatGPT diagnosed a child’s condition as a branchial cleft cyst instead of the correct Branchio-oto-renal syndrome, a more complex genetic condition that also involves the formation of branchial cleft cysts.

The study exposed a notable deficiency in ChatGPT-4’s diagnostic approach. The AI struggled to recognize known relationships between conditions, a skill crucial for medical diagnosis. In one case, it failed to link autism with scurvy (Vitamin C deficiency), a common issue due to dietary restrictions common in neuropsychiatric conditions. This oversight is critical, as neuropsychiatric conditions are known risk factors for vitamin deficiencies in children, especially in high-income countries.

Pediatric diagnoses require special consideration of the patient’s age, a factor that adds complexity, particularly when dealing with infants and small children who cannot explain their symptoms clearly. The study’s findings emphasize the indispensable role of clinical experience, suggesting that human pediatricians are far from being replaced by AI in diagnostics.

Introducing ‘Link History,’ Facebook’s Ominous Tech for Tracking Your Every Move

Gizmodo reports that Facebook’s new feature, “Link History,” is the latest addition to its vast library of data collection tools. This new feature automatically logs every link a user clicks within the Facebook app, storing it in a dedicated section in the app. While there is an opt-out option for users, the default setting is to collect this data which will mainly be used for enhancing targeted advertising.

While tech regulations are becoming more strict and companies such as Apple and Google are stepping up their privacy game, Mark Zuckerberg’s Meta ( the parent company of Facebook and Instagram) seems to be continuing its usual plan of action to collect as much information on its users as possible. This move appears to be a strategic effort to preserve its data-collection focused business model in the face of growing privacy concerns and regulatory challenges.

Facebook is marketing Link History as a user convenience feature, a tool for users to keep track of their browsing activities in one place. The feature is being advertised as a means for users to “never lose a link again.” However, Meta makes it clear that enabling Link History allows the company to leverage this information to fine-tune ad targeting on its various platforms.

Google Section

Google Pays Parents to Collect Facial Recognition Data of Kids

Google has reportedly launched a new data collection project, hiring a contractor to gather facial recognition data from children, offering $50 to parents for their child’s participation.

404 Media reports that the project revolves around Google collecting specific data, including eyelid shape and skin tone of children. Parents are asked to film their children wearing various props, such as hats and sunglasses. TELUS, acting on Google’s behalf, is paying parents $50 for this data collection effort. The primary objective of this project is to build datasets for machine learning, artificial intelligence, and facial recognition technologies.

This method of data collection is a drastic shift from the traditional approaches of scraping existing images or analyzing previously collected materials found on the internet. Instead, Google is directly engaging the public in its data gathering processes and compensating them for their contributions. Google clarified that this initiative is part of their efforts to verify users’ age.

Participants in the project are required to be minors, specifically aged between 13 and 17 years. The process involves taking multiple short videos, each under 40 seconds, with the overall task expected to take 30 to 45 minutes. The project insists that the filming of children be done on private premises, not in public spaces, and requires parental consent.

TELUS, which also offers facial recognition products, stated that the purpose of this collection is to capture a broad cross-section of participants. This aims to ensure that the services and products derived from this data are representative of a diverse set of end-users. The data collection is intended to improve authentication methods, thereby offering more secure tools for users.

Google Claims Its ‘Robot Constitution’ Will Prevent AI Harm

The Verge reports that the DeepMind robotics team at Google has recently revealed three significant advancements aimed at enhancing the decision-making capabilities of robots, particularly in terms of safety and efficiency. One of these achievements is the development of a Robot Constitution, a new approach to AI safety inspired by Isaac Asimov’s famous Three Laws of Robotics.

This constitution is a set of safety-focused guidelines designed to instruct large language models (LLMs) in making decisions that consciously avoid tasks involving humans, animals, sharp objects, and electrical appliances.
...
An interesting aspect of this initiative is Google’s AutoRT data gathering system, which integrates a visual language model (VLM) with the LLM. This combination allows the AI to understand and adapt to its environment, making appropriate decisions based on the context. According to the company, this has helped to enhance the AI’s ability to operate autonomously in various settings.

In practical terms, the Robot Constitution includes several safety mechanisms, such as automatic shutdown if the force on the robot’s joints exceeds a specified threshold and a physical kill switch for human operators. Over seven months, Google conducted over 77,000 trials with a fleet of 53 AutoRT robots in different office environments, testing various degrees of autonomy and control.

The Controversial Elon Musk Section

Elon Musk’s Tesla Cuts Range Estimates on Multiple Models

The Verge reports that in a recent update, Tesla has adjusted the range estimates for several of its electric vehicles, most notably the Model Y, Model S, and Model X. This change has led to a decrease in the advertised range by up to 37 miles, bringing its performance estimates more in line with the experiences of existing owners.

The Model Y Performance now shows a range of 285 miles, down from 303, and the Model Y Long Range is listed at 310 miles, a decrease from 330. The Model X Plaid sees a smaller reduction, going from 333 miles to 326. The biggest cut can be seen with the Model S Plaid with 19-inch wheels, dropping from 396 miles to 359 miles.

Interestingly, this change in range estimates currently only applies to Tesla’s U.S. market, with no similar adjustments reported for the company’s UK and EU listings. Other Tesla models, including the Cybertruck, Model 3, and the Model Y Rear-Wheel Drive, are unaffected by this update.

Tesla has not explicitly stated the reasons for these reduced range estimates, but the change is likely due to a combination of factors, such as the company’s continuous improvements in “comfort and functionality” that may consume more energy, and possibly new EPA testing requirements that could result in increased energy consumption and a slight dip in the range.

NLRB Claims SpaceX Illegally Fired Employees for Criticizing Elon Musk

The New York Times reports that the National Labor Relations Board (NLRB) has made serious allegations against SpaceX, the space transport services company led by CEO Elon Musk. In a recent filing, the NLRB claims that SpaceX wrongfully terminated eight employees who had been vocal in their criticism of Musk’s public conduct. These employees had distributed a letter internally, which pointedly criticized Musk’s behavior on social media, particularly focusing on his handling and response to sexual harassment allegations against him.

The letter called for the company to take a clearer and more consistent stance on harassment issues. It specifically requested the enforcement of policies that would prevent similar incidents and foster a safer, more respectful work environment. However, instead of addressing these concerns, SpaceX is accused of retaliating against the employees involved by terminating their employment. This move, according to the NLRB, is a violation of federal labor laws that protect employees’ rights to engage in concerted activities for mutual aid and protection.

Paige Holland-Thielen, one of the employees who was fired, stated: “At SpaceX the rockets may be reusable, but the people who build them are treated as expendable. I am hopeful these charges will hold SpaceX and its leadership accountable for their long history of mistreating workers and stifling discourse.”

The complaint filed by the NLRB details further alleged misconduct by SpaceX executives. Notably, it includes an instance involving President Gwynne Shotwell, who is said to have intervened to stop the spread of the letter criticizing Musk.

This legal challenge against SpaceX is not an isolated incident in Musk’s corporate empire. One of Musk’s other companies, Tesla, has been previously sanctioned by the labor board for illegally firing an employee involved in union organizing. Musk’s recent acquisition of Twitter and subsequent mass layoffs, including the termination of individuals who publicly disagreed with him, have caused many to accuse Musk of being unable to handle dissent within his companies.

Tesla Recalls 1.6 Million Electric Vehicles Sold in China over Autopilot Crashes

The automaker will be able to remotely fix the vehicles with software upgrades to come into compliance with the China State Administration for Market Regulation, reports Reuters.

The recall includes some imported Model S, Model X and Model 3 vehicles, as well as Model 3 and Model Y cars produced in China between August 2014 and December 2022.

The Chinese regulator reported that Tesla drivers may have misused their vehicles’ autopilot features, increasing dangerous collisions and other safety risks. The product recall, to be rolled out in multiple stages, mirrors the U.S. recall announced just last month.

As Breitbart News reported on December 13, American safety regulators issued a statement to alert Tesla customers that Elon Musk’s company would be recalling nearly all vehicles sold in the United States.

Upwards of two million units across the automaker’s model lineup will be impacted by that recall, which also plans to address defective software to ensure drivers are correctly using the autopilot function.

The United States recall came after two years of investigation by the National Highway Traffic Safety Administration (NHTSA) into a series of crashes that occurred while the car was in the partially-automated driving mode.