Honda may want to ax the plug. 

The automaker filed a patent application for a “contactless electric power transmission system.” Honda’s filing describes a system for rapid wireless charging for electric vehicles: A battery attached to the underside of the vehicle lets it charge on the go.

Honda’s invention is basically a Mario Kart Dash Panel in real life: While a vehicle is in motion, if it passes over the contactless power transmission system built into the ground, the vehicle will gain back some of its charge.

In order to prevent excess power from being transmitted, and to ensure that the power transmission system is ready for the passing vehicle, the car and the system are wirelessly connected prior to the vehicle reaching the “transmission zone” to communicate the vehicle’s power needs. That way, the vehicle still benefits from the charge even if it’s traveling at a high speed. 

Honda’s patent also noted that, when this system communicates with the vehicle, it may pass along billing information associated with a user’s account to pay for the charge, similar to an automatic toll booth. 

Charging infrastructure remains one of the biggest hurdles for widespread EV adoption. While charging stations are widely available in many cities, many hesitate to pull the trigger due to “range anxiety,” said Rei Vardi, CEO and founder of EV rental and sharing firm Eon. 

But range isn’t the only issue, Vardi noted. The speed and efficiency of charging also present a barrier. As it stands, even the fastest chargers take around 20 minutes to get an electric vehicle to a full charge, he said, and many chargers can take several hours to power up a vehicle. 

“You don’t want to find yourself in a situation where you’re close to zero and you don’t really know where to charge,” Vardi said. “Alternatively, maybe there’s a bunch of charging stations nearby, but each one of those is going to take five to six to 12 hours.” 

Honda has been ramping up EV production in North America — it’s planned an $11 billion investment in Canadian EV operations and is retooling its Ohio plant to manufacture EVs — so it makes sense why the firm would also be researching new innovations in charging infrastructure that may help cure range anxiety. 

“The expansion of EV charging infrastructure and the introduction of more advanced EV charging technologies could help bolster EV adoption by reducing some of the lingering concerns from potential EV buyers,” said Madeline Ruid, AVP and research analyst covering clean energy at Global X ETFs.

This invention, however, might not be it, said Vardi. Contactless charging relies on close proximity between the device and the charging element. Given that vehicles are at least a foot or so off the ground, that distance between the power transmission system and the battery would make it difficult for the vehicle to benefit in any way. 

Plus, contactless charging often takes longer than plug-in — so the milliseconds it’d take for the vehicle to drive over this charger likely wouldn’t provide enough time to do much of anything. “The viability is really not there,” said Vardi. “The inefficiency, comparatively, with contactless charging is so much greater.”

Zoom wants to make sure you’re closing. 

The company filed a patent application for “engagement analysis between groups of participants.” Zoom’s tech relies on a trained model to pick out which meeting participants spoke the most, and who stayed on mute. 

The filing specifically notes this tech’s usefulness in the context of sales teams, as they “often dissect and analyze such sales meetings with prospective customers after they are conducted,” Zoom said in the filing. “Within a sales context, of particular importance is whether or not the customer or prospective customer is ‘engaged’ during the meeting.”

After a meeting ends, Zoom’s tech transcribes the audio and extracts individual phrases spoken by each participant. The system then groups those phrases into two distinct speaker groups: each representing one organization, team, or company.

The system then calculates engagement metrics for each speaker within those groups, measuring things like amount of times spoken and speech duration, questions asked, response times, and topic relevance. Those metrics are translated into an engagement score for both the overall meeting and each participant. 

It adds up that Zoom would seek to patent such tech. As AI has taken the tech industry by storm, the company has spent the past year and a half launching machine learning and AI-powered productivity tools, including an AI collaboration platform called Zoom Workplace and a note-taking tool called Zoom Docs. 

But a productivity tool like the one this patent describes often walks a fine line when it comes to employee privacy, said Phil Libin, co-founder and CEO of AI startup studio All Turtles and video presentation app mmhmm. While tech like this could be useful in the context of employee training, it could easily verge on invasive depending on who the engagement metrics are presented to. 

For example, if those metrics are given over to the employee who took part in the meeting to improve their performance, then Zoom’s tech could be greatly beneficial. However, if these metrics are handed over to supervisors — or used to automate processes such as performance reviews — then it could be cause for concern. “I could see it going in both directions,” said Libin. 

The way it goes could depend on whether or not a tool like this is created with “beneficence,” said Libin, or in the best interest of a fully-informed user. “If a user was fully educated and informed about what this was doing and why it was doing it, would the user want this to be done?” he said. “I think that’s what determines whether something is worthwhile.”

Microsoft wants to make AI models more accurate. 

The company filed a patent application for “enriching language model input with contextual data.” This basically helps Microsoft’s model offer more accurate responses and predictions by feeding it domain-specific data as part of the input. 

“Existing [natural language processing]-based technologies are incomplete or inaccurate,” Microsoft said in the filing. “One reason is because existing language models, such as Large Language Models (LLM), often make predictions without adequate contextual data.” 

To give its language models a bit more context, Microsoft’s tech introduces “corpus data supplements,” or extra information that’s specific to the needs of the organization, as inputs for the model. These supplements can include metadata or tags that help distinguish general information from specific knowledge. 

For example, if a project is called “Red Sea,” and a person asks their AI assistant, “What is the status of Red Sea?” that term may be tagged in Microsoft’s system as “[project name: business unit X],” to make sure it doesn’t answer questions about the body of water instead. Microsoft noted that this tech makes AI more accurate and wastes less computational resources on unnecessary queries. 

With Microsoft’s primary use case for AI being a co-pilot to supercharge productivity, it adds up that it’d look for ways to make its outputs more relevant and accurate for whatever company or organization is using it. This also isn’t the first time that Microsoft has sought to use context and extra data to make its AI better. 

Additionally, the company claims it has found a way to fix when its AI models do make missteps. Last week, the company unveiled a tool called “correction,” which attempts to catch and fix when AI-generated text is inaccurate by comparing it to source material. The feature can be used with the output of any large language model. 

“Empowering our customers to both understand and take action on ungrounded content and hallucinations is crucial, especially as the demand for reliability and accuracy in AI-generated content continues to rise,” Microsoft said in its announcement.

Microsoft isn’t the only one that’s attempted to figure out what to do when AI models hallucinate. Google unveiled a similar tool earlier this year, called Vertex AI. 

The problem, however, is that eliminating hallucination from AI is nearly impossible. Given that Microsoft’s tech uses AI to help fix AI, it’s possible that those models, too, could hallucinate, causing more harm than good. One expert told TechCrunch that hallucination is “an essential component of how the technology works.”