🧑‍🎓 Speak to Me Easy…AI Says…Clever Than You, ChatGPT Aces Med, Law, & Bizniz

In this edition: 🧑‍🎓 Speak to Me Easy…AI Says…Clever Than You, ChatGPT Aces Med, Law, & Bizniz

Research highlights: Speak Easy: Human Language Proves a Powerful Teaching Tool for Robots

The value of words as teaching aids for robots has been demonstrated. How can you train a robot the most effectively? Sometimes all you need to do is talk openly and clearly to it. Researchers discovered that employing human-language descriptions of tools can hasten a simulated robotic arm's learning of lifting and handling various tools. The learning of a simulated robotic arm lifting and utilizing multiple tools can be sped up by employing human-language descriptions of the items, according to Princeton researchers who are investigating novel ways to train robots.

The findings add to the body of research showing that giving autonomous robots richer information during artificial intelligence (AI) training might increase their capacity for situational adaptation, hence enhancing their efficacy and safety.

The robot's proficiency in manipulating new tools that weren't part of the first training set was enhanced by include descriptions of a tool's form and function in the training procedure. The novel approach, called Accelerated Learning of Tool Manipulation with LAnguage, or ATLA, was presented on December 14 at the Conference on Robot Learning by a group of mechanical engineers and computer scientists with the title Leveraging language for accelerated learning of tool manipulation.

Industry news: ChatGPT- The Robot Scholar, Outperforming Human Students on Exams

Recently, ChatGPT received a passing grade in:

• The US Medical Licensing Exam, a three-part test

• A Wharton Business School examination—the final exam for the Operations Management course in the MBA program

• Exams for the Constitutional Law, Employee Benefits, Taxation, and Torts courses at the University of Minnesota

There were short-answer, essay, and multiple-choice questions on these assessments, and ChatGPT passed each one with passing marks but not perfect scores. Examiners remarked that, like most people, ChatGPT performed significantly better in writing than in math, most people are, haha.

It's possible that forcing robots to take a ton of tests will force them to rebel against humanity. The highly publicized AI text generator ChatGPT from OpenAI passed exams in law, commerce, and medicine this month and performed better than many human students.

Policy and regulation: Consumers Strike Back: EU Strengthens Product Liability for AI and Software

The EU's strict and tort product liability for flawed AI and software is increasing. AI, software, and even digital services in the IoT sector are included under the current iteration of the Product Liability Directive 85/374/EEC as "products" that are subject to strict product liability.

The change broadens the list of parties who might be held accountable to some fulfillment service providers and online marketplaces. Consumers will be able to seek compensation from non-EU manufacturer representatives thanks to the updated Product Liability Directive. Unsettlingly, the draft eliminates the current strict liability liability cap. There are various procedural changes in the revision: It provides a type of discovery in which culpable businesses will be obligated to provide claimants with the evidence they will need to support their claims in court.

The burden of proof has been changed, and now the claimant may only need to show a reasonable probability that the injury was brought on by the goods due to the product's high level of technological and scientific complexity. By shifting the burden of proof, the companion draft of an AI Responsibility Directive seeks to apply the same principles to product liability under national tort laws.

AI and society: Predicting the Future with AI: A Double-Edged Sword for Society's Progress

AI might impede society progress.

History demonstrates that daring actions by outliers—individuals willing to propose novel ways of understanding the world—often lead to progress in society. The majority of AI systems forecast the future using historical data. The capacity of outliers to alter the rules of the game and improve society may be constrained when AI plays a larger role in decision-making.