Discover more from PATENT DROP
PATENT DROP: Palantir watches over your shoulder
Plus: Ford charges responsibly; Twilio hushes your data
Happy Thursday and welcome to Patent Drop!
Here’s some important news: Patent Drop has a new home.
Earlier this year we announced The Daily Upside acquired this newsletter and starting next week, Patent Drop is moving to a new email sender.
Here’s what you should do to continue getting this newsletter without disruption:
Add firstname.lastname@example.org to your Contacts or VIPs.
And starting with next week, if you don’t see the newsletter in your inbox on Thursday, search for email@example.com and drag the newsletter to your Primary inbox if it’s in Spam or Promotions.
Now let’s go into today’s newsletter!
Today, we dive into Palantir’s tech to keep a closer eye on employees; a filing from Ford to charge its EVs at the best times; and a patent from Twilio to keep your voice chats under wraps.
Palantir spies on your sales calls
Palantir wants to give its clients a panopticon for their workforce.
The company filed a patent application for a system for optimized calculation of “member activity metrics” — members meaning employees, in this case. Palantir’s tech offers a way for administrators or managers to perform “complex analyses” of employee performance, as well as compare members within a cohort.
Palantir’s system is essentially an automated, data-backed HR manager: It keeps a close eye by hoarding tons of employee activity data, such as “emails, phone calls, messages, contacts, or meetings,” picking out the relevant data points to calculate activity metrics for each employee. These metrics are then displayed on an interactive user portal that allows administrators to run reports on user metrics and compare employees to one another, such as those with similar experience, similar positions or similar locations.
The interface includes visual aids, such as different kinds of graphs, and formatting to make the metrics easier to understand, Palantir notes. Reports can be broken up by type of activity and timeframe, and the system lets managers know when an employee performance review is due.
Palantir says that current available employee metrics systems don’t work well for large organizations, and lack the ability to scale. Existing tools also “lack the ability to quickly traverse an organization's data of varying sources and formats (e.g., emails, voice messages) to find and analyze data associated with members,” Palantir noted.
If we’re being frank, Palantir doesn’t have a reputation of trust and transparency in the public eye. The Peter Thiel-founded company is better known for its work with government agencies, licensing its so-called spy tech with the capability to track and gather data on just about anything you can think of.
That said, Palantir looking into new ways to extrapolate information from data is hardly surprising. The company’s bread and butter is data analytics, working with a number of both enterprise customers and government agencies to help them dig up data and understand what it means — and a product like this could be appealing to both camps, said Cornelio Ash, director and equity analyst at William O’Neil.
“At the core of it, they're a data analytics company. The end goal is making better decisions faster and more accurately from (analyzing) a large data set.,” said Ash. “They're very focused on large enterprise customers. But anyone that has a large dataset and needs to get better information out of it would benefit from Palantir.”
Big companies and government organizations with large workforces — especially those with remote employees — are the likely target for this kind of tech. But the employees monitored under this system likely wouldn’t take too kindly to it, especially as employee mistrust in leadership is worsening. According to a recent Gallup poll, only 21% of employees strongly agree that they trust the leadership of the organizations they work for.
The lack of trust seemingly goes both ways, given that the demand for employee surveillance software has increased 54% since March 2020.
“I don't think it's popular. It's kind of like a big brother thing,” Ash said. “You're tracking employees at a very intricate level. But I think companies want to do that, especially amid this kind of hybrid work environment. You want to know who's doing what, where, how efficient they're being, and if they're being honest or transparent about their workload.”
Ford’s thoughtful charger
Ford wants to make charging its EV fleet more energy-conscious.
The company wants to patent a system for an EV “charging control system” which takes into account predicted charging demand on the utility grid during certain times to lessen stress and cut down on energy expenses.
Ford’s system first monitors how much the grid is relying on “reactive power” — the energy that keeps the grid running smoothly, but isn’t doing the heavy lifting — in a given moment. The system monitors a metric called “power factor,” which determines how much of that reactive power the EV is taking up as it charges. The system then only charges the vehicle from the grid if the use of that reactive power is below a certain threshold, and if the power factor metric is within a specific range.
If Ford’s system determines that it shouldn’t charge using the power grid, it will use the vehicles within its fleet to charge each other using bi-directional transfers of power. In layman's terms, this means that a battery from one car can charge the battery of another car under certain circumstances.
For example, say one car’s battery in a fleet is at 15%, while another car’s battery is at 85%. The car with the higher battery percentage can charge the one with the lower battery percentage so they can both be above a certain threshold.
Ford’s argument for this system is basically that sharing is caring. “ While this depletes a portion of the energy of the first battery, it may be used to increase the charge of a second battery, such that both batteries are above a predetermined threshold,” the company said.
While the impact of reactive power on consumer EV users is relatively low, the same can’t be said for enterprise- or industrial-grade operations. Given that Ford offers a line of EV utility vehicles, adding a method for distribution enterprises to charge them at a lower cost could be a lucrative opportunity, said Matt McCaffree, VP of utility market development at clean energy parking and mobility company FLASH.
Distribution centers specifically are an “interesting challenge to solve for,” said McCaffree. “Those fleets … are typically driving under 200 miles a day, and are really great targets for electric vehicles and electrification.”
Ford’s solution to balancing power between utility vehicles is novel in that no other company or organization has “quite cracked this nut yet,” said McCaffree. However, a lot of companies have been researching how to solve the problem of keeping large fleets of EVs charged in this way, he said, so solutions may be sprawling and “shouldn’t be limited to one patent.”
Though every automotive company has thrown their hat in the EV ring over the years, Ford has been especially ambitious. The company wants to manufacture EVs at a run rate of more than 2 million by the end of 2026 and has filed to patent a number of EV-related inventions. Ford electrifying one of its most popular vehicle models, the F-150, in and of itself shows “how serious they are about this and how big of a player they will be in the future,” said McCaffree.
Getting into industrial and utility EVs with this patent only further proves the case that wide-scale electrification of its lineup is more than just a pet project for the company.
Twilio’s magic eraser
Twilio wants to make sure your data is hush-hush – literally.
The communications tech company is seeking to patent a system for “personal information redaction and voice deidentification.” As the title implies, this tech uses the concept of data anonymization, or obfuscating customers’ identities in audio clips, to allow audio from conversations to be reviewed without sacrificing customer privacy.
“Protecting privacy of people is an important concern, so before the data is used for business or government purposes, there may be a need for the data to be anonymized to enable the use of the data without compromising privacy,” Twilio said in its filing.
Twilio’s system works in two parts: First, it identifies and redacts any form of personally identifiable information within an audio clip. (this can be anything from a full name to an address to a social security number.) Using a machine learning algorithm trained on audio data, transcript data and redacted transcripts, the system automatically picks through audio data to obscure any personal or sensitive information in the clip and replaces it with “beeps or silence,” without a reviewer needing to analyze the transcript themselves after the fact.
Twilio’s system then takes it a step further by modifying the entire clip itself to change the voice of the customer to a “neutral voice” to avoid “the possibility that the voice of the user may be recognized.”
Along with keeping personal data safe, Twilio notes that the anonymized data could be used for a number of business solutions, including using it to train other AI algorithms without putting customer privacy at risk.
Taking a less-is-more approach is generally one of the best ways to protect customer privacy, because you can’t lose what you don’t have. Twilio seems to be taking this approach: By replacing and redacting sensitive information from its audio data altogether, the company is removing the possibility of customer exposure via security breach from the equation all together.
Data minimization, however, isn’t exactly new. Plenty of companies use this technique to cover themselves in the event of a breach. We’ve seen patents for similar tech, too: PayPal filed to patent a system to mask sensitive information in unstructured data (photos, video or audio); and Ford is seeking to patent a way to anonymize speech data by getting rid of “speaker-identifying characteristics.” While Twilio’s use case is somewhat narrow, it’s uncertain the company will be able to patent this tech, said Ali Allage, CEO and president of BlueSteel Cybersecurity.
Still, this method of cybersecurity is particularly important when it comes to audio data, as it’s especially difficult to keep safe compared to other kinds of data, said Allage. For one, because unstructured data is more difficult to store, it’s more difficult to protect.
Plus, while messaging platforms often have some form of encryption that users are made aware of, said Allage, audio-based interactions often don’t offer that same assurance. “You have an operator telling you that a call is being recorded, and then it's capturing everything,” said Allage. “There is no encryption or decryption methodology that's being told to the end user.”
Given that Twilio’s entire business relies on the ability for its enterprise customers – which include major companies like Lyft, Airbnb and Dell – to safely communicate, keeping their data under lock and key is crucial. But the company has had issues in the past with data security, suffering two major security breaches from the “0ktapus” hacker group last summer. Using this patent’s methods could minimize the severity if breaches occur in the future.
However, the patent leaves something to be desired in regards to ownership of information by Twilio’s customers, said Allage. While the data may be protected by Twilio, he said, the big question is “How is all this (data) stored, and who holds the keys of encryption and decryption?” he said.
Some other fun patents we wanted to share.
Apple wants to make sure you feel represented. The company wants to patent a system for “camera-less representation of users” that relies on animated avatars (read: FaceTime Memojis).
Meta wants to make gaming a more personal experience. The company filed a patent application for an AR gaming system which overlays games within video communications.
Adobe wants to make color more accessible. The company filed a patent application for a system that creates a “color texture map specific to a given color vision deficiency.”
What else is new?
AI firm Hugging Face raised $235 million at a valuation of $4.5 billion. The company received backing from the biggest names in tech, including Google, Amazon, Nvidia, Salesforce and more.
Nvidia soared past Wall Street’s expectations in its latest earnings report, with shares opening at a record high this morning.
Meta is launching its own generative AI coding assistant, called Code Llama, to compete with Microsoft, OpenAI and Google.