• Copilot is Microsoft's generative AI tool now featured in much of its productivity programs.
• Copilot can draft text, analyze information and data, and suggest ideas.
• Copilot has also been plagued with security issues, and customers have criticized its effectiveness.

The entire tech industry is mired in an AI arms race, and Microsoft bet big on Copilot, its generative artificial intelligence chatbot.

Microsoft released Copilot in 2023, and rapidly rolled it out across various products and softwares. The company markets Copilot as a tool to help users with productivity tasks such as drafting a memo for work, adding to or amending hectic calendars, analyzing a spreadsheet or a few lines of code, or even writing a poem or short story.

By combing the vastness of the internet in fractions of a second to source troves of information and then engaging in machine learning and informed prediction, Copilot can create content; it can analyze, interpret, and explain extant data; and it can create systems for planning and managing many aspects of your life, from work to recreation to hobbies and more.

"We believe Copilot will be the new UI that helps you gain access to the world's knowledge and your organization's knowledge, but most importantly, it will be your agent that helps you act on that knowledge," Microsoft CEO Satya Nadella said during a keynote address at Microsoft's annual Microsoft Ignite business conference in November 2023.

Not long after Copilot's launch, industry experts predicted that for the fiscal year 2024, Copilot would generate billions for the company. However, the AI tool is not without its flaws. Customers and company insiders have criticized Copilot for its ineffectiveness and cost, and IT leaders have questioned its value to their companies; the widespread disappointment in Copilot has raised doubts over its long-term profitability.

What is Microsoft Copilot AI?

Microsoft Copilot is an AI-powered chatbot that relies on large language models (often shorted to LLMs) to help users with productivity and content creation tasks. The more you use it, the more it learns about your interests, preferences, and habits, and the better it tailors itself to serving your needs. It's available for use on Windows, Macs, and both the Apple iOS, and Android mobile platforms.

You can use Microsoft Copilot in many programs that you likely already use, including Microsoft Teams, Microsoft Word, Microsoft Excel, Microsoft PowerPoint, OneNote, Outlook, and more. It can assist you in rapidly creating and editing documents, spreadsheets, presentations, and more, acting both as your content creator and an editor and sort of AI coworker.

The basic version of Microsoft Copilot is free to use on the web, in Windows, with a Mac OS, and with Android and iOS. The free version includes access to GPT-4, GPT-4 Turbo, and GPT-4o during non-peak times. It also allows users to create and edit AI images, use plugins, and more.

However, there is also a paid version of Copilot called Copilot Pro, which offers more advanced features and better access. Copilot Pro includes all the features of the free version, plus priority model access and the ability to use Copilot in Microsoft apps like Word, Excel, PowerPoint, Outlook, and OneNote.

To use Copilot Pro, you need a Microsoft 365 Personal or Family subscription, which costs $6.99 per month for a personal subscription and $9.99 per month for a family subscription, good for up to six users. With Copilot Pro added, the total cost is around $26.99 per month.

Is Microsoft Copilot better than ChatGPT?

In some ways and for some uses, Copilot can be more effective than ChatGPT. Copilot is better for quick, tactical tasks that come about during your workflow, while ChatGPT is more commonly used for broader tasks and conversational AI, like for writing a creative story or rehearsing before a meeting or interview.

Copilot is part of the Microsoft ecosystem and can easily pull information from across Microsoft applications, which can be helpful, and it can also reply to questions with visual responses, such as photos and images. Also, unlike ChatGPT, which doesn't provide sources for its responses, Copilot does, making cross-referencing and fact-checking easier. That said, the paid version of ChatGPT is only $20 a month, so it's $7 cheaper than Copilot Pro.

What are the risks of using Microsoft Copilot?

Using Microsoft Copilot can pose several risks, including data leakage. Customers have already raised multiple security issues with Copilot, and in some cases delayed deployment over the concerns.

Copilot can generate outputs that include sensitive data, which may be shared with the wrong audience as the AI simply doesn't know better. For example, while you use Microsoft Teams, Copilot could summarize conversations and record action items against your better judgement, which could accidentally reveal private information.

There is also the all-too-common human problem of over-reliance. As Copilot and other AI tools become increasingly integrated into daily life, users may unknowingly rely on it more, which could impact investment and strategic decisions, how they write and express themselves, and frankly how much creative and critical thinking they do.

If you asked someone what they wanted from TVs released in 2025, I doubt they'd say "more software and AI." Yet, if you look at what TV companies have planned for this year, which is being primarily promoted at the CES technology trade show in Las Vegas this week, software and AI are where much of the focus is.

The trend reveals the implications of TV brands increasingly viewing themselves as software rather than hardware companies, with their products being customer data rather than TV sets. This points to an alarming future for smart TVs, where even premium models sought after for top-end image quality and hardware capabilities are stuffed with unwanted gimmicks.

LG’s remote regression

LG has long made some of the best—and most expensive—TVs available. Its OLED lineup, in particular, has appealed to people who use their TVs to watch Blu-rays, enjoy HDR, and the like. However, some features that LG is introducing to high-end TVs this year seem to better serve LG’s business interests than those users' needs.

Read full article

Comments

• I had the rare opportunity to fly in a historic jet trainer ahead of the Pacific Airshow in October.
• The T-33 Shooting Star was used to train US Air Force pilots on jet aircraft until the early 1960s.
• The flight was an unforgettable glimpse into aviation history as modern fighters rule the skies.

Strapped in the rear seat of the T-33's cockpit, I triple-checked the buckles and straps securing me. The clear bubble canopy sealed shut with a click, shutting out the aircraft's engine roar.

As someone who has solely flown on commercial airliners and once on a puddle jumper in Alaska, I didn't know what to expect until the jet's mounting speed on the runway pressed me back into my seat. The rush of adrenaline pulled my focus away from the sensation of the wheels leaving the ground as the world below transformed into a patchwork of shrinking buildings and roads.

As the sunset painted wispy clouds in a red-orange glow, I felt transported back to the Jet Age's earliest days, when competency at this aircraft's stick and analog controls was a make-or-break for US Air Force trainees.

Developed by Lockheed Corporation, the T-33 Shooting Star was a subsonic jet trainer designed to transition Air Force pilots from propeller-driven aircraft to the emerging era of jet propulsion in the late 1940s.

But the hourlong flight in the run-up to the Pacific Airshow in Huntington Beach, California, taught me to appreciate the iconic aircraft that paved the way for the cutting-edge jets that dominate today's skies.

America's first jet trainer

The US was among several countries around the world ushering in a new era of fighter aircraft toward the end of World War II.

The T-33 Shooting Star was instrumental in transitioning early military aviators — then already qualified to fly propeller-driven aircraft — to fighter jets. The jet trainer was a direct offshoot of the single-seat P-80 Shooting Star, one of the first operational jet fighters flown by the US Air Force, adding another seat to the cockpit to accommodate both student and instructor.

Equipped with a turbojet engine, the T-33 could reach speeds of up to 600 mph and altitudes of up to 45,000 feet. Just shy of 40 feet long, the jet trainer had a large clear bubble canopy, providing ample visibility to the pilots in the two-seat cockpit decked out with various flight indicators and engine gauges.

Lockheed manufactured over 5,000 so-called "T-Birds" between 1948 and 1959, used by air forces around the world until the fleet was slowly phased out starting in the 1980s. The US Air Force retired its last T-33, a variant modified for aerodynamic research, in 1997, ending the storied aircraft's five-decade tenure with the service.

While the T-Bird no longer trains military pilots, the aircraft's legacy has been preserved in aerial museums and heritage flights.

Seasoned airshow pilot

I flew with Greg "Wired" Colyer, a longtime performer at the Pacific Airshow. Colyer founded Ace Maker Aviation, operating three T-33 aircraft for aerial demonstrations like airshows.

A US Army veteran and aviation enthusiast, Colyer completed his fixed-wing training while studying at the US Navy's Naval Postgraduate School in Monterey, California.

He told Business Insider he wanted to fly military helicopters until a knee injury prompted him to instead work as an air traffic controller for the Federal Aviation Administration for nearly three decades.

In 2008, Colyer restored his first T-Bird, aptly naming the historic jet trainer the "Ace Maker," teaching himself aerobatic maneuvers for airshows across the US. Aside from airshows, the California native also works as a test pilot instructor for the US Navy and Air Force.

Over the next decade, he added two more T-33s, the Ace Maker II and Ace Maker III, the latter of which I flew in.

"It was just one of my favorite planes as a kid," Colyer said. "It was our very first jet trainer and just an iconic piece of US history."

During the Pacific Airshow, Colyer put the T-33's capabilities on full display, dazzling spectators below with his self-taught aerobatic maneuvers.

The airshow also featured performances from the F-22 Raptor, a fifth-generation stealth jet considered to be the top US air superiority fighter. The Thunderbirds, the Air Force's aerial demonstration squadron, also wowed crowds at the airshow, flying in tight formations at near-supersonic speeds in F-16 Fighting Falcon aircraft.

Inside the cockpit

After haphazardly climbing atop the T-33's wing, I peered into the cockpit and was immediately intimidated by the crowded display of gauges and controls.

As a Cold War-era aircraft, the T-33 didn't feature the digital displays or avionics that are routine on more advanced fighter jets. The T-Bird had primarily analog controls, leaving the pilot responsible for manually controlling the plane's speed, altitude, and flight path.

Colyer walked me through a crash course of the T-33's rear instrument panels — the airspeed indicator, altimeter, turn coordinator, heading indicator, radio controls, and various fuel and engine gauges monitoring temperature, pressure, and fuel state.

As a tandem trainer, the backseat also had a yoke and rudder pedals, allowing student pilots to practice controls and maneuvers. I was explicitly instructed not to mess with either.

Though rudimentary to an experienced pilot, I could see how it would benefit early military aviators to develop essential skills for operating the advanced jet aircraft of the time.

As a journalist, however, the mechanics of the aircraft were lost on me, but I paid close attention to two controls in the cockpit:

1. The switch that muted my transceiver so I wouldn't distract the pilot.
2. The lever that operated the plane's rocket-powered emergency ejection seat — for obvious reasons.

Preparing for takeoff

As Colyer strapped into the pilot seat, I settled in the seat behind him, my head barely peeking over the edge of the cockpit.

I adjusted my flight helmet and oxygen mask, snapping a few shameless selfies to show off in group chats with my friends and family after the flight.

Colyer's voice crackled in my headset as he communicated with the control tower. In an instant, the plane's engine roared to life as the plane taxied the airfield in Orange County.

A thrilling takeoff

With one final check-in, I signaled my OK with a meager thumbs-up to the pilot before we sped down the runway and into the air.

The acceleration pinned me against the seat, and I peered at the blurring silhouettes of buildings and aircraft surrounding us, slowly shrinking as we gained altitude.

Panoramic views

I had to stifle my gasp — lest I distract Colyer while he operated the aircraft — as we were soon graced with an unobstructed view of the skies through the glass canopy.

Experiencing the T-33 in action

While Colyer spared me from the thrills — and terrors — of complex airshow maneuvers, I was still acutely aware of the G-forces pressing down on me with every sharp movement of the aircraft.

A slight nudge of the yoke sent us banking into a turn, the horizon tilting dramatically. I was repeatedly startled by urgent beeping from the control panel, only to be reassured by Colyer that it was because the plane was idling to maintain slower speeds.

He said the T-33 was only going around 250 mph — a fraction of what the jet trainer was capable of and nearly half the speed he normally goes during airshow performances.

Trainers flying in tandem

We were accompanied by a former US Navy Texan T-6 during the flight demonstration. A fellow trainer aircraft, the T-6 is a two-seat jet trainer produced by Raytheon Aircraft in the 1990s and early 2000s.

The T-6 was primarily used to train students in basic flying skills through the Air Force and Navy's entry-level training program known as the Joint Primary Aircraft Training System.

T-33 touchdown

Time seemed to pass more slowly as we flew for about an hour, constrained by the fuel capacity of the T-33, which burns about 300 gallons per hour.

Descending back toward the runway, the adrenaline began to ebb before Colyer informed me that he would partially pop open the canopy to create additional drag upon landing. Lowering the visor on my flight helmet, I braced myself against rushing winds as the T-33 touched down with a thud and rolled to a halt on the tarmac.

The sky is the limit

After sharing the cockpit selfies and views with friends and family, I fielded a number of questions about the flight and whether I got to deploy any onboard weapons (trainer aircraft are not armed). It was a chance to reflect on the unforgettable experience.

In short, we flew in circles over Orange County, cruising at speeds slower than commercial passenger planes. I then spent the rest of the time ahead of the Pacific Airshow talking with demo pilots who operated more advanced frontline fighters.

But the experience was more than just a golden-hour flight in a vintage military jet trainer. The intense rumble of the turbojet engine, the array of analog gauges and controls, and sweeping views through the canopy offered a glimpse into what it must have felt like for early military aviators mastering jet-powered flight for the first time.

While I'm still holding out on future familiarization flight rides on modern fighter jets, my flight on the T-Bird will serve as a constant reminder of just how far US military aviation has come.

In the business world, there are few areas that artificial intelligence hasn't touched. Many industries are rushing to adopt AI, and the technology is changing how employees collaborate and complete tasks.

Generative AI is a major buzzword for business leaders. But actually integrating AI can be a different story.

"A lot of our clients have dozens of AI pilots everywhere," Jack Azagury, the group chief executive for consulting at Accenture, said at one Workforce Innovation roundtable. "Very few have a coherent business case and a true reinvention and transformation."

How do companies move forward as the novelty of AI wears off? Business Insider's Julia Hood asked members of the Workforce Innovation board how they transitioned their AI pilots into real-world use cases. Board members shared five major ways their companies were moving AI from theory to operations.

"Before we go and tell our clients to embark on AI fully, we want to be an AI-first organization," said Anant Adya, an executive vice president, service-offering head, and head of Americas delivery at Infosys. "We want to show our clients we are using AI, whether it is in HR when it comes to driving better employee experience or when it comes to recruitment."

Members also highlighted employee training and peer-to-peer learning opportunities.

The roundtable participants were:

• Anant Adya, an executive vice president, service-offering head, and head of Americas Delivery at Infosys.
• Lucrecia Borgonovo, a chief talent and organizational-effectiveness officer at Mastercard.
• Neil Murray, the CEO of Work Dynamics at JLL.
• Justina Nixon-Saintil, a vice president and chief impact officer at IBM.
• Marjorie Powell, a chief HR officer and senior vice president at AARP.

The following has been edited for length and clarity.

Identify early adopters, like human resources

Nixon-Saintil: Because we provide these platforms and solutions to clients, we are usually client zero. We implemented AI across our business and multiple functions, and one of the first things we did was our AskHR product, which I think answered over 94% of questions employees had.

HR employees now spend time doing higher-order work and partnerships with business units instead of answering basic questions that a virtual assistant can answer. I think that's when you start seeing a lot of the benefits of it.

Borgonovo: HR has been leading the way in terms of embedding AI to enhance the employee experience end to end, right before you hire somebody all the way to after they leave the organization. There are tons of opportunities to improve performance and productivity and provide greater personalization.

Invest in ongoing training

Adya: There are certain AI certifications and courses that everybody has to take to be knowledgeable about AI. So we are driving education in terms of what is the impact of AI, what is gen AI, what are LLMs, and how you look at use cases. And certainly educating everybody that it's not about job losses but about amplifying your potential to do more.

Powell: We have hands-on skill building. This past year we posted over 20 AI workshops helping teams integrate AI into their work. We really encourage our staff to participate. We have a product we're using behind our firewall, so they can engage and play with it. We're just telling them go ahead and try to break it, so they can give us feedback on what's working.

There was a team of people who said we want to see how you could use AI with PowerPoint or Excel. And they're finding, well, it's not so good in those things. But as it continues to grow, they'll be ready for that, and they'll know what it was able to do and what it wasn't. I think it's just making it fun, and that way it's not so scary.

Murray: Our internal large language model is now a widget on everybody's dashboard that is accessible on your landing page. Training is super important here to make people comfortable with it. Even if it's just an online module, you have to get people comfortable.

Nixon-Saintil: We've also done companywide upskilling. We had two Watsonx challenges. Watsonx is our AI data platform. This is one of the ways we've upskilled a majority of the organization. The outcome of that is there are some great ideas that employees actually ideated, and they're now implementing those ideas and solutions in different functions.

Borgonovo: Employees want to use AI, and I think they're eager to learn how to use AI to augment their jobs. For that, we built a three-tiered learning approach. One is democratizing access for everybody and building general knowledge of AI.

The second tier is much more role-specific. How do we drive new ways of working by having people in different roles embrace AI tools? Software engineering, consulting, sales — you name it. And then something we definitely want to build for the future is thinking proactively about how you re-skill people whose roles may be impacted by AI so they can become more comfortable doing high-level tasks or can shift to a different type of role that is emerging within the organization.

The other piece is where we're seeing the greatest demand internally, which is for knowledge management. It's gathering information from a lot of different sources in a very easy way.

Another job family that is very eager to get their hands on new AI technology is software engineering. We have taken a very measured approach in deploying coding assistants within the software-engineering community. This year we did a pilot with a subset of them using coding assistants. The idea is to just learn and, based on our learning, scale more broadly across our software-engineering community in 2025.

One of the really interesting learnings from this pilot was that the software engineers who were using the coding assistants probably the best were people who had received training. What we're learning is that before you start rolling out all of these technologies or AI-specific platforms for different job families, you have got to be really intentional about incorporating prompt training.

Unlock peer-to-peer learning

Powell: We have idea pitch competitions and a year-round idea pipeline program where people can put in ideas on how to use AI and share what they've learned. It sparks a lot of peer learning and creativity on our digital-first capabilities to help us with our digital transformation.

Then we collaborate through community. We have a generative-AI community of practice. This is somewhat like how companies have employee resource groups; we have communities of practice as well. They give employees a space to share their techniques and learn from each other and stay ahead of evolving trends. They meet monthly, they have an executive sponsor, and they have all kinds of activities and learning opportunities.

Murray: As we monitored AI use and what sort of questions were being asked, we identified super users across all departments — so the people who were capable of developing the most evolved prompts. I suppose those prompts are now appearing in pull-down menus to help people who maybe aren't as advanced in their use of it, because prompting is a really important part of this. And so the super users are driving everybody else to show them what's possible across the organization.

Find customer pain points to solve

Borgonovo: One of the use cases that drives not only knowledge management but also efficiencies is around customer support. Customer support is probably one of the areas that has been leading the way.

We have a customer onboarding process that can be very lengthy, very technical, involving hundreds of pages of documentation and reference materials. It was our first use case for a chat-based assistant that we processed in terms of streamlining and creating greater efficiency and a much better customer experience.

Reinforce responsible leadership

Powell: We want our leaders, people leaders particularly, to guide employees to use AI effectively and responsibly. We want to make sure they're emphasizing privacy, policy, and efficiency. So we encourage managers to point the staff toward training that we offer, and we offer quite a bit of training.

Microsoft-owned GitHub announced on Wednesday a free version of its popular Copilot code completion/AI pair programming tool, which will also now ship by default with Microsoft’s popular VS Code editor. Until now, most developers had to pay a monthly fee, starting at $10 per month, with only verified students, teachers, and open source maintainers getting […]

© 2024 TechCrunch. All rights reserved. For personal use only.

• I recently drove the Honda Pilot and Hyundai Palisade midsize, three-row family SUVs.
• The Palisade delivers a refined experience with top-notch tech content worthy of a luxury brand.
• The Pilot's smartly designed cabin, easy-to-use tech, and superb quality are great for families.

American consumers purchase more than 2 million midsize family SUVs every year.

It's a broad segment in which brands can take vastly different approaches.

I recently had the chance to drive two popular three-row midsize SUVs, the Honda Pilot and Hyundai Palisade, that found success by appealing to the needs of different families.

The Honda Pilot is a thoroughly practical family hauler with a smooth V6 engine, easy-to-use tech, and a thoughtfully designed cabin that provides loads of space for people and cargo.

The Hyundai Palisade, on the other hand, delivers effective daily transportation with a luxurious flair rarely found in a mass-market family SUV.

Here's a closer look at how the two stack up against one another.

Which is a better deal?

The base front-wheel-drive Pilot Sport starts at $39,900, while the entry-level, front-wheel-drive Palisade SE starts at $36,650.

My well-equipped, Alabama-built Pilot Elite AWD test car started at $52,780 with an as-tested price of $54,630, while my loaded Hyundai Palisade Calligraphy AWD starts at $52,100. With fees, the Ulsan, South Korea-built SUV's as-tested price is $53,650.

Which one looks better?

The Pilot's blunted, truck-like front end looks great, but the Palisade's parametric jewel-style front grille takes the cake for me. It's a front facia that's both futuristic and elegant.

Which one is larger?

The Honda Pilot is one of the largest SUVs in its class. At 200 inches, my test car is about three and a half inches longer than the Hyundai Palisade.

It's also taller and wider than the Palisade.

The battle of the V6s

While many of their contemporaries have turned to small, turbocharged four-cylinder engines for motivation, the Pilot and Palisade are both powered by smooth, naturally aspirated V6s.

The Palisade's 3.8-liter V6 produces 291 horsepower and 262 lb-ft of torque. It's paired with an 8-speed automatic transmission.

The Pilot is powered by a 3.5 liter, 285 horsepower V6 with 262 lb-ft of torque paired with a 10-speed automatic transmission.

To help save fuel, both engines are equipped with the universally loathed automatic stop-start system, while the Honda engine can also shut down three of its six cylinders when high power output isn't needed.

The EPA rates both vehicles for 19 mpg city and 21 mpg combined driving. However, the Pilot gets 25 mpg of fuel economy on the highway, one more than the Palisade.

Which has a better all-wheel-drive system?

The Pilot and Palisade are both naturally front-wheel-drive crossover SUVs with all-wheel-drive available as an optional extra.

Honda's i-VTM4 and Hyundai's HTRAC systems both operate in front-wheel drive when cruising, sending power to the back only when it detects traction loss.

What sets Honda's system apart is that it can target how much power each of the rear wheels receives, such as distributing more power to the outside rear wheel in a corner to reduce understeer and improve controllability.

Which is better to drive?

Neither vehicle is particularly sporty to drive. The Pilot and Palisade lack the instant low-end torque of a turbocharged powerplant or an electric motor. The duo is also saddled with transmissions tuned for efficiency over outright performance.

Even though both models returned identical 0 to 60 mph runs of 6.9 seconds in Motor Trend's testing, the Palisade inched ahead of the Pilot in this department as it provided a more refined and pleasant driving experience.

The Palisade's drivetrain delivered power effortlessly, while it felt like the Pilot's engine and transmission combo had to strain a bit to get the vehicle going.

Which has a nicer cabin?

While both the Pilot and Palisade feature spacious, smartly designed cabins with high-quality materials and excellent ergonomics, they were executed in very different ways.

The Palisade's cabin is stylish and refined. It stands out for its quiet, isolating ride, design features like the airplane seat-style wing-out headrests, and creature comforts like the massaging ergo-motion driver seat.

The Pilot shines for its practicality and thoughtful touches, which help it excel at hauling people and their stuff. One such touch is the ingenious removable middle seat module, which can quickly transform the second-row captain's chairs into bench seats to accommodate extra passengers.

Which has more passenger space?

The Pilot and Palisade have two of the roomiest passenger cabins in the segment.

The Palisade offers about an inch and a half more second-row legroom compared to the Pilot, but it trails the Pilot by an inch in third-row legroom.

Which has the best tech?

The Pilot's standard seven-inch and optional nine-inch displays are verging on unacceptably small for a vehicle in this segment. Especially when compared to the Palisade's standard 12.3-inch infotainment touchscreen.

Both systems are well-sorted, easy to use, and backed up by physical switches to help make use on the go a bit easier. However, I found the Hyundai system to be a bit more intuitive to use.

Apple CarPlay and Android Auto compatibility come standard on both vehicles.

Both vehicles also come with an intercom system that allows the driver and rear cabin passengers to communicate through the SUVs' speakers.

How much cargo can they haul?

The Honda Pilot is ahead of the Palisade in overall cargo space, with up to 112.1 cubic feet of cargo space behind the first row versus 86.4 cubic feet for the Palisade.

My Verdict: It's a tie.

The Honda Pilot and Hyundai Palisade represent excellence in the genre but through very different means.

The Palisade's level of luxury and refinement is truly noteworthy. I wouldn't feel cheated if Hyundai slapped and Genesis badge on it and jacked up the price by $10,000.

On the other hand, the Pilot's quality and thoughtful execution are the mark of a brand with a quarter century of experience building minivans that can withstand years of grueling school runs and the destructiveness of a bored, curious child.

Given their vastly different personalities, the vehicle I'd choose depends purely on my family's changing needs.

With younger children, I'd choose the Pilot's roomy, easy-to-use, and durable cabin. But as they and I grow older, count me in on the Palisade's refined, isolating ride and heated massaging seats.

Microsoft’s AI can now read your screen — or rather, the websites you’re browsing. On Thursday, the company began rolling out a limited, U.S.-only preview of Copilot Vision, a tool that can understand and respond to questions about sites you’re visiting using Microsoft Edge. Gated behind Copilot Labs, an opt-in program for experimental AI capabilities, […]

© 2024 TechCrunch. All rights reserved. For personal use only.

• Microsoft's Bing search engine has Copilot AI features built into it.
• Copilot, integrated into Bing, can perform tasks like writing poems and making reservations.
• Here's how to use the AI service in Bing.

Microsoft first unveiled a revamped, AI-powered version of its search engine, Bing, last year.

The new Bing, which promised to be "more powerful than ChatGPT," runs on Microsoft's own next-generation language model called "Prometheus." The proprietary technology was developed using elements of OpenAI's most advanced GPT models as part of Microsoft's partnership with the company.

Built into the revamped search engine is Microsoft's AI chatbot, Copilot, which can perform a number of tasks the old Bing never dreamed of, like suggesting recipes, writing poems, conducting image-based search queries, and making restaurant reservations. Copilot was formerly called Bing Chat.

Microsoft's Bing was launched in 2009, more than a decade after Google's launch — and though it's come a long way since then, Bing still holds a fraction of the market share compared to Google.

And though Microsoft CEO Satya Nadella once raved that Copilot would "fundamentally transform our relationship with technology," the AI assistant has struggled to live up to the hype, both inside and outside the company.

Here's a step-by-step walkthrough of how to access and use the new Bing with Copilot.

How to use the new Bing

1. Go to Bing.com/new in your internet browser.

   Note: You don't need to download Microsoft's web browser, Edge, but if you do, Copilot is integrated directly into the browser, with an icon in the top-right corner that lets you chat with the AI companion.

   

2. There are a few ways to use AI in your search experience from the Bing homepage.

3. One option is to click "Try now" underneath the heading "Bing generative search," located directly below the general search bar on the Bing homepage.

   

   • The "Try now" button takes you to a search results page that auto-populates for the query "How can I get started with learning to play the guitar."
   • The page displays several helpful resources related to that query, including a Table of Contents with sections that guide you through the process of learning to play guitar, alongside videos and step-by-step instructions.

   

   • You can also type in your own query, like "How to paint a bathroom," and, depending on the query you search, the results will offer helpful sections including relevant videos, instructions, and, in this case, a section on the side for the best paint to use on bathroom cabinets.

     

4. Another option is to click "Copilot" in the bar at the very top of the Bing homepage.

   

   • This takes you to the Copilot homepage where you can message the chatbot.

   • In the "Message Copilot" text field at the bottom of the page, you can type something you need help with, like, for example, "Write me a poem," or "plants that survive with minimal light." The chatbot will quickly give you a detailed response — for example, it offered 10 options for plants that don't need much light.

     

Google got some disappointing news at a status conference Tuesday, where US District Judge Amit Mehta suggested that Google's AI products may be restricted as an appropriate remedy following the government's win in the search monopoly trial.

According to Law360, Mehta said that "the recent emergence of AI products that are intended to mimic the functionality of search engines" is rapidly shifting the search market. Because the judge is now weighing preventive measures to combat Google's anticompetitive behavior, the judge wants to hear much more about how each side views AI's role in Google's search empire during the remedies stage of litigation than he did during the search trial.

"AI and the integration of AI is only going to play a much larger role, it seems to me, in the remedy phase than it did in the liability phase," Mehta said. "Is that because of the remedies being requested? Perhaps. But is it also potentially because the market that we have all been discussing has shifted?"

Read full article

Comments

• Elon Musk criticized the F-35 and called crewed fighters obsolete in the drone era.
• Musk's comments align with tech leaders advocating for drones over traditional military assets.
• Drones can't yet replace crewed aircraft. Even if they could, a mix of both might be more effective.

Drones are changing war in ways we never thought possible, but are we to the point where uncrewed systems can replace top-dollar weapons like the F-35 stealth fighter?

Prominent tech industry figures are saying yes. Former warfighters and analysts say we aren't there yet, and replacement might not be the right call regardless.

Tesla CEO Elon Musk has targeted the Pentagon's prized fifth-generation stealth jet, the F-35 Lightning II Joint Strike Fighter. In a series of social media posts on X this week, he called it idiotic to continue building them and criticized the design. Pointing to Ukraine, he said human-piloted jets are "obsolete" and "inefficient" and will "just get pilots killed" as drones and counter-air threats become more prolific.

In the Ukraine war, drones are surveilling and striking enemy vehicles and troop positions. But they are not a substitute for crewed jets, which Kyiv has long sought in greater numbers even as pilots face a tough air-defense environment.

Musk's comments follow similar remarks by former Google CEO Eric Schmidt, who called tanks "useless" last month while urging the Army to "give them away" and "buy a drone instead." Musk went a bit further, speculating about ways adversaries could defeat the F-35's stealth.

Musk's criticism comes as he prepares to target wasteful government spending as part of Trump's Department of Government Efficiency initiative. The F-35 Joint Strike Fighter is the Pentagon's most expensive weapon system program, with lifetime costs expected to top $2 trillion. Musk has previously suggested the F-35, troubled by setbacks throughout its development, isn't the best fit for the military.

Four years ago, the SpaceX founder said a remotely controlled uncrewed fighter would be a better alternative to the F-35 and argued the future is autonomous drone warfare.

This week, he said that "manned fighter jets are obsolete in the age of drones."

Drones are game-changers

Small, cheap drones are transforming land warfare by providing new options for tactical reconnaissance, targeting solutions, and threatening maneuvers. For situations where air and sea combat over vast areas might be more prevalent, like a war in the US military's priority Indo-Pacific theater, these drones are too slow with inadequate payloads and range to be sufficient.

The US also needs fast, low-observable, and maneuverable platforms that can carry advanced sensors and stand-off weaponry across great distances through potentially contested airspace.

"That's just not something that small UAVs can do," said Justin Bronk, a Royal United Services Institute airpower analyst.

Providing the full range of capabilities for this theater means larger, more sophisticated platforms with a higher price tag. Existing remotely controlled systems only meet some of the demands, some can cost as much as an F-35, and they are vulnerable to intensifying electronic warfare and surface-to-air threats.

The US military is actively developing new semi-autonomous and artificial intelligence-driven aircraft, from pilotless F-16s to collaborative combat aircraft in which a pilot directs the tasks. This space offers immense potential, with some limitations since the technology isn't yet mature.

"If I develop an aircraft that does not require a human in the cockpit, I could develop one that could pull 15 Gs, 20 Gs because you're no longer worried about the physiology of the human," said Guy Snodgrass, a retired naval aviator and former senior defense official.

Without a human pilot, "you could then strip out the cockpit, you could strip out the oxygen generation, you could strip out a lot of the life support systems," potentially freeing up space for sensors, weapons, and more, the former TOPGUN instructor said, arguing that "there are definitely advantages."

But without crewed fighter aircraft, particularly the high-end systems like the F-35, the US risks being "stuck with a huge capability gap for a significant period of time because the drone technology and the ability to not only produce it but then to incorporate it in the military and actually employ it in a tactically relevant or strategically relevant sense isn't there yet," he said.

Mixing the crewed fighters with uncrewed aircraft

In response to Musk's comments on X about their fighter this week, a spokesperson for Lockheed Martin told BI that the F-35 Joint Strike Fighter is "the most advanced, survivable, and connected fighter aircraft in the world, a vital deterrent and the cornerstone of joint all-domain operations," a reference to the jet's role as a combat quarterback.

The fifth-gen stealth jet is not only a US military aircraft. It is used by nations around the world, with more planning to establish F-35 fleets. That's because the F-35 isn't just a fighter jet. It's also a bomber, electronic warfare plane, surveillance tool, battle management platform, and key communications node.

An uncrewed aircraft can't yet match that capability. "That technology is simply not there," said Mark Gunzinger, a retired US Air Force pilot and the director of Future Concepts and Capability Assessments at the Mitchell Institute for Aerospace Studies.

If one day drones have all those capabilities, there are still advantages to keeping human pilots flying combat missions. Combat is unpredictable and requires making decisions in uncertain situations. Autonomous systems might be less prepared to look past the data, like a false radar return, to make the smart call.

Machines are more rigid. "The flexibility that human pilots give you to use the machine and the systems that it has in relatively unforeseen circumstances or across a very wide variety of mission types and circumstances is something that's very difficult to replicate in an automatic system," Bronk said.

For the US military, wargaming scenarios have shown that the better solution is not one or the other, crewed or uncrewed.

"We need both," Gunzinger said. "And the greatest impact on warfighting, the biggest leap ahead in warfighting capabilities and capacity, is in figuring out how to combine what they both bring to the fight in the most effective way. That's the secret sauce."

Much like the US isn't solely dependent on a single energy source, US national security isn't dependent on a single capability. In this situation, uncrewed systems enhance crewed systems and vice versa. The US military is still figuring out what comes next, but the F-35 is an imperfect but important bridge to that future of airpower, whether it's crewed, uncrewed, or some mix of both.

Over five months after publicly scrapping the first version of the Windows Recall feature for its first wave of Copilot+ PCs, Microsoft announced today that a newly rearchitected version of Recall is finally ready for public consumption.

For now, the preview will be limited to a tiny subset of PCs: Qualcomm Snapdragon X Elite and Plus Copilot+ PCs enrolled in the Dev channel of the Windows Insider program (the build of Windows that includes Recall is 26120.2415). Intel and AMD Copilot+ PCs can’t access the Recall preview yet, and regular Windows 11 PCs won’t support the feature at all.

If you haven’t been following along, Recall is one of Microsoft’s many AI-driven Windows features exclusive to Copilot+ PCs, which come with a built-in neural processing unit (NPU) capable of running AI and machine learning workloads locally on your device rather than in the cloud. When enabled, Recall runs in the background constantly, taking screenshots of all your activity and saving both the screenshots and OCR’d text to a searchable database so that users can retrace their steps later.

Read full article

Comments

• Microsoft released tools to address security issues with its AI assistant Copilot.
• Copilot's indexing of internal data led to oversharing of sensitive company information.
• Some corporate customers delayed Copilot deployment due to security and oversharing concerns.

You know when a colleague overshares at work? It's awkward at best.

Microsoft's Copilot has been doing an AI version of this behavior, which has unnerved corporate customers so much that some have delayed deploying the product, as Business Insider first reported last week.

Now, the software giant is trying to fix the problem. On Tuesday, Microsoft released new tools and a guide to help customers mitigate a Copilot security issue that inadvertently let employees access sensitive information, such as CEO emails and HR documents.

These updates are designed "to identify and mitigate oversharing and ongoing governance concerns," the company explained in a new blueprint for Microsoft's 365 productivity software suite.

"Many data governance challenges in the context of AI were not caused by AI's arrival," a Microsoft spokesperson told BI on Wednesday.

AI is simply the latest call to action for enterprises to take proactive management of their internal documents and other information, the spokesman added.

These decisions are controlled by each company's unique situation. Factors such as specific industry regulations and varying risk tolerance should inform these decisions, according to the Microsoft spokesperson. For instance, different employees should have access to different types of files, workspaces, and other resources.

"Microsoft is helping customers enhance their central governance of identities and permissions, to help organizations continuously update and manage these fundamental controls," the spokesman said.

Copilot's magic — its ability to create a 10-slide road-mapping presentation, or to summon up a list of your company's most profitable products — works by browsing and indexing all of your company's internal information, like the web crawlers used by search engines.

Historically, IT departments at some companies have set up lax permissions for who can access internal documents — selecting "allow all," say, for the company's HR software, rather than going through the trouble of selecting specific users.

That never created much of a problem, because there wasn't a tool that an average employee could use to identify and retrieve sensitive company documents — until Copilot.

As a result, some customers have deployed Copilot, only to discover that it can enable employees to read an executive's inbox or access sensitive HR documents.

"Now, when Joe Blow logs into an account and kicks off Copilot, they can see everything," said one Microsoft employee familiar with customer complaints. "All of a sudden Joe Blow can see the CEO's emails."

Are you a Microsoft employee or someone else with insight to share?

Contact Ashley Stewart via the encrypted messaging app Signal (+1-425-344-8242) or email ([email protected]). Use a nonwork device.

In the short-term, the most dangerous thing about AI language models may be their ability to emotionally manipulate humans if not carefully conditioned. The world saw its first taste of that potential danger in February 2023 with the launch of Bing Chat, now called Microsoft Copilot.

During its early testing period, the temperamental chatbot gave the world a preview of an "unhinged" version of OpenAI's GPT-4 prior to its official release. Sydney's sometimes uncensored and "emotional" nature (including use of emojis) arguably gave the world its first large-scale encounter with a truly manipulative AI system. The launch set off alarm bells in the AI alignment community and served as fuel for prominent warning letters about AI dangers.

On November 19 at 4 pm Eastern (1 pm Pacific), Ars Technica Senior AI Reporter Benj Edwards will host a livestream conversation on YouTube with independent AI researcher Simon Willison that will explore the impact and fallout of the 2023 fiasco. We're calling it "Bing Chat: Our First Encounter with Manipulative AI."

Read full article

Comments