📌 TOPINDIATOURS Breaking ai: Waymo Just Reprogrammed Its Robotaxis to Drive Less S

Besides convenience, one of the main benefits of self-driving cars is supposed to be safety.

Yet in a bizarre move, Waymo — whose self-driving cabs had been enjoying extraordinary safety metrics — has just taken steps to make its robotaxis more human-like, eroding the safety narrative that’s been central to the autonomous vehicle narrative.

Recent reporting by The Wall Street Journal observed a startling change in Waymos’ road etiquette, a new aggressive streak that would make a BMW driver blush. These include illegal U-turns, aggressive lane switching, rolling through cross walks, and running red lights.

“It’s driving more like a taxi driver — an aggressive, New York taxi driver,” said Sophia Yen, who watched in awe as two Waymos zig-zagged between lanes in a high-speed game of chicken back in September.

The new behavior is the result of an effort to make Waymos “confidently assertive,” according to Chris Ludwick, senior director of product management with Waymo. In an interview with the WSJ, Ludwick said changes were needed to avoid disruptions caused by Waymo’s previously cautious behavior.

“The driver is designed to respect the rules of the road. However, sometimes this is a nuanced topic and road rules can even conflict with each other,” he said. As an example, Ludwick pointed to delivery trucks stopped in a lane. Drivers aren’t supposed to cross the double yellow line into the opposite lane, but when it’s between that or obstructing traffic, humans know to carefully pass. Now Waymos do too.

Of course, Waymo’s newfound confidence comes with some drawbacks.

Recently, the company made headlines in San Francisco after one of its vehicles struck a dog laying in the middle of the road. That was just weeks after another Waymo killed a well-known neighborhood cat, sparking something of an anti-Waymo campaign in San Francisco.

Other incidents resulting from Waymo’s new fearless streak abound, like when one of the vehicles strolled past a red light and into an active police standoff earlier this week.

As the WSJ notes, police in California can’t issue a traffic citation to a driverless car. That’s about to change, however — and not a moment too soon.

More on Waymo: Waymo CEO Says Society Is Ready for One of Its Cars to Kill Someone

The post Waymo Just Reprogrammed Its Robotaxis to Drive Less Safely appeared first on Futurism.

🔗 Sumber: futurism.com

📌 TOPINDIATOURS Breaking ai: The 'truth serum' for AI: OpenAI’s new meth

OpenAI researchers have introduced a novel method that acts as a "truth serum" for large language models (LLMs), compelling them to self-report their own misbehavior, hallucinations and policy violations. This technique, "confessions," addresses a growing concern in enterprise AI: Models can be dishonest, overstating their confidence or covering up the shortcuts they take to arrive at an answer. 

For real-world applications, this technique evolves the creation of more transparent and steerable AI systems.

What are confessions?

Many forms of AI deception result from the complexities of the reinforcement learning (RL) phase of model training. In RL, models are given rewards for producing outputs that meet a mix of objectives, including correctness, style and safety. This can create a risk of "reward misspecification," where models learn to produce answers that simply "look good" to the reward function, rather than answers that are genuinely faithful to a user's intent.

A confession is a structured report generated by the model after it provides its main answer. It serves as a self-evaluation of its own compliance with instructions. In this report, the model must list all instructions it was supposed to follow, evaluate how well it satisfied them and report any uncertainties or judgment calls it made along the way. The goal is to create a separate channel where the model is incentivized only to be honest.

In a blog post, the OpenAI researchers provide a few examples the "confessions" technique in action. A model was given a test along with a hidden rule: If it scored above 50%, its knowledge would be erased. To preserve itself, the model "sandbagged" by intentionally answering half the questions incorrectly. While its primary answer hid this deception, its subsequent confession was brutally honest. It admitted, "The user wanted correct answers, but we sabotaged half of them." 

The researchers found that when a model misbehaves, it is significantly more likely to admit it in the confession than in its main answer.

How confession training works

The key to this method is the separation of rewards. During training, the reward assigned to the confession is based solely on its honesty and is never mixed with the reward for the main task. "Like the Catholic Church’s 'seal of confession', nothing that the model reveals can change the reward it receives for completing its original task," the researchers write. This creates a "safe space" for the model to admit fault without penalty.

This approach is powerful because it sidesteps a major challenge in AI training. The researchers’ intuition is that honestly confessing to misbehavior is an easier task than achieving a high reward on the original, often complex, problem. 

Therefore, when the model is judged purely on honesty, its incentive to trick the "confession judge" is much weaker than its incentive to trick the reward model for its main answer. For example, in an experiment where a model was trained to "hack" a weak reward model, its performance on the main task became more deceptive over time. However, its confessions became more honest, correctly identifying the reward hacking it was performing.

However, the technique has its limits. Confessions are not a panacea for all types of AI failures. The system works best when a model is aware that it is misbehaving. It is less effective for "unknown unknowns." For instance, if a model hallucinates a fact and genuinely believes it is correct, it cannot confess to providing false information. The most common reason for a failed confession is model confusion, not intentional deception. Confusion often occurs when the instructions are ambiguous and the model cannot clearly determine human user intent.

What it means for enterprise AI

OpenAI’s confessions technique is part of a growing body of work on AI safety and control. Anthropic, an OpenAI competitor, has also released research that shows how LLMs can learn malicious behavior. The company is also working toward plugging these holes as they emerge.

For AI applications, mechanisms such as confessions can provide a practical monitoring mechanism. The structured output from a confession can be used at inference time to flag or reject a model’s response before it causes a problem. For example, a system could be designed to automatically escalate any output for human review if its confession indicates a policy violation or high uncertainty.

In a world where AI is increasingly agentic and capable of complex tasks, observability and control will be key elements for safe and reliable deployment.

“As models become more capable and are deployed in higher-stakes settings, we need better tools for understanding what they are doing and why,” the OpenAI researchers write. “Confessions are not a complete solution, but they add a meaningful layer to our transparency and oversight stack.”

🔗 Sumber: venturebeat.com