Connect with us

Startups

AI Weekly: Watson Health and semi-autonomous driving failures show the dangers of overpromising AI

Published

on

This week in AI, reality came knocking — both for AI health tech and semi-autonomous driving systems. IBM agreed to sell the assets from its Watson Health business to investment firm Francisco Partners, following the division’s steeply declined performance. Meanwhile, the Insurance Institute for Highway Safety (IIHS), a nonprofit financed by the insurance industry, announced a new ratings program designed to evaluate how well “partial” automation systems like Tesla’s Autopilot provide protections to prevent misuse.

The twin developments are emblematic of the AI industry’s perennial problem: acknowledging AI’s limitations. Slate’s Jeffrey Funk and Gary Smith do a thorough job of recapping overly optimistic AI predictions in recent years, including Ray Kurzweil’s pronouncement that computers will have “human-level” intelligence and “the ability to tell a joke, to be funny, to be romantic, to be loving, to be sexy” by 2029.

As any expert will confirm, AI is nowhere close to approaching human-level intelligence — emotional or otherwise. (Kurzweil’s new estimate is 2045.) Similarly, autonomous cars and AI-driven health care haven’t achieved the lofty heights futurists once envisioned they would reach. It’s an important lesson in setting expectations — the future isn’t easy to predict, after all — but also an example of how the pursuit of profit supercharges the hype cycle. Under pressure to show ROI, some health tech and autonomous vehicle companies have collapsed under the weight of their overpromises, as this week’s news shows.

Riding high on the Watson win against Jeopardy! champion Ken Jennings, IBM launched Watson Health in 2015, positioning the suite of AI-powersed services as the future of augmented care. The company’s sales pitch was that Watson Health could analyze reams of medical data — far faster than any human doctor could, ostensibly — to generate insights that improve health outcomes.

IBM reportedly spent $4 billion beefing up its Watson Health division with acquisitions, but the technology proved to be inefficient at best — and harmful at worst. A STAT report found that the platform often gave poor and unsafe cancer treatment advice because Watson Health’s models were trained using erroneous, synthetic medical records rather than real patient data.

Watson Health’s demise can be partially attributed to IBM CEO Arvind Krishna’s shifting priorities, but growing cynicism about AI’s health capabilities no doubt also played a role. Studies have shown that nearly all eye disease datasets come from patients in North America, Europe, and China, meaning eye disease-diagnosing algorithms are less certain to work well for racial groups from underrepresented countries. An audit of a UnitedHealth Group algorithm determined that it could underestimate by half the number of Black patients in need of greater care. And a growing body of work suggests that skin cancer-detecting algorithms tend to be less precise when used on Black patients, in part because AI models are trained mostly on images of light-skinned patients.

Semi-autonomous, AI-powered driving systems are coming under similar scrutiny, particularly as automakers ramp up the rollout of products they claim can nearly drive a car themselves. In October 2021, Tesla was ordered to turn over data to the National Highway Traffic Safety Administration as part of an investigation into the company’s cars crashing into parked vehicles. The suspicion was that Tesla’s Autopilot was responsible — either in part or whole — for the dangerous behaviors.

It’s not an unreasonable assumption. Late last year, Tesla rolled out an update to Autopilot with a bug that caused the automatic braking system in Teslas to engage for no apparent reason. This caused cars to rapidly decelerate as they traveled down the highway, putting them at risk of being rear-ended.

Tesla isn’t the only vendor that’s struggled to perfect semi-autonomous car technology. A sobering 2020 study from the American Automobile Association found that most semi-autonomous systems on the market — including those from Kia and BMW — ran into problems an average of every eight miles. For example, when encountering a disabled vehicle, the systems caused a collision 66% of the time.

In 2016, GM was forced to push back the rollout of its Super Cruise feature due to unspecified issues. Ford recently delayed its BlueCruise system in order to “simplify” the tech.

That brings us to this week’s news: the Insurance Institute’s ratings program to evaluate semi-autonomous systems’ safety protections. The group hopes it’ll encourage automakers to create better designs once the first set of ratings, currently in development, are issued this year.

“The way many of these systems operate gives people the impression that they’re capable of doing more than they really are,” Insurance Institute research scientist Alexandra Mueller said in a statement. “But even when drivers understand the limitations of partial automation, their minds can still wander. As humans, it’s harder for us to remain vigilant when we’re watching and waiting for a problem to occur than it is when we’re doing all the driving ourselves.”

That’s all to say that AI — whether self-driving or disease-diagnosing — is fallible, just like the humans who design it. Visions of a Jetsons future might be tantalizing, but when lives are at stake, history has shown that it’s best to be overly cautions.

For AI coverage, send news tips to Kyle Wiggers — and be sure to subscribe to the AI Weekly newsletter and bookmark our AI channel, The Machine.

Thanks for reading,

Kyle Wiggers

Senior Staff Writer

Source link

Startups

8 Ways You Can Save Yourself and Others From Being Scammed

Published

on

Opinions expressed by Entrepreneur contributors are their own.

Statistics on the number of scam websites that litter the internet are disturbing. During 2020, Google registered more than 2 million phishing websites alone. That means more than 5,000 new phishing sites popped up every day — not to mention the ones that avoided Google’s detection. In 2021, the U.S. Federal Bureau of Investigation (FBI) reported nearly $7 billion in losses from cybercrime that is perpetrated through these sites.

What exactly are scam websites? Scam websites refer to any illegitimate website that is used to deceive users into fraud or malicious attacks. Many scammers operate these fake websites and will download viruses onto your computer or steal passwords or other personal information.

Reporting these sites as they are encountered is an important part of fighting back. In other words, if you see something, say something. Keeping quiet, even if you avoid falling prey, allows the scammers to aim at another target.

Perhaps you’ve received a suspicious link in an email? Or maybe a strange text message that you haven’t clicked on. Fortunately, there are many organizations out there that have launched efforts aimed at reducing the threat that they pose. In general, these organizations put scam websites on the radar by collecting and sharing information about them. In some cases, they prompt an investigation into the scammers behind the sites.

Related: Learn How to Protect Your Business From Cybercrime

It’s free to report a suspicious website you’ve encountered, and it takes just a minute. Here are eight ways you can report a suspected scam website to stop cyber criminals and protect yourself and others online.

1. The Internet Crime Complaint Center

The IC3, as it is known, is an office of the FBI that receives complaints from those who have been the victims of internet-related crime. The IC3 defines the internet crimes that it addresses to include illegal activity involving websites. Complaints filed with the IC3 are reviewed and researched by trained FBI analysts.

2. Cybersecurity and Infrastructure Security Agency

CISA, which is an agency of the U.S. Department of Homeland Security, targets a wide range of malicious cyber activity. It specifically requests reports on phishing activity utilizing fraudulent websites. Information provided to CISA is shared with the Anti-Phishing Working Group, a non-profit focused on reducing the impact of phishing-related fraud around the world.

3. econsumer.gov

The econsumer.gov site, run by the International Consumer Protection and Enforcement Network, is for reporting international scams. It is supported by consumer protection agencies and related offices in more than 65 countries. A secure version of their site is used by law enforcement agencies to share info on scams.

4. Google Safe Browsing

While Google does not have a mechanism for reporting all varieties of website scams, there is a form for reporting sites that are suspected of being used to carry out phishing. Reports made via the form are managed by Google’s Safe Browsing team. Google’s Transparency Report provides information on the sites that it has determined to be “currently dangerous to visit.”

Related: Is That Instagram Email a Phishing Attack? Now You Can Find Out.

5. PhishTank

This service was founded by Cisco Talos Intelligence Group to “pour sunshine on some of the dark alleys of the Internet.” Phishtank includes an ever-growing list of URLs reported as being involved in phishing scams. To date, it has received more than 7.5 million reports of potential phishing sites. It says that more than 100,000 of the sites are still online.

Related: 6 Ways Better Business Bureau Accreditation Can Boost Your Business

6. Antivirus Apps

Antivirus providers such as Norton, Kaspersky, and McAfee have forms that can be used to identify pages that users feel should be blocked. Scam sites would definitely fall under that category. With some antivirus platforms, reporting forms can only be accessed by registered users. Norton’s is open to anyone.

7. Web host

There is a chance that the DNS service hosting the scam site will take action to shut it down. There are a variety of online resources that can help you to find the DNS of a particular site. Once you identify it, send a message to their customer service reporting the site in question and the experience that you had.

8. Share your experience on social media

This is actually more like sounding an alarm than filing a report, but it might protect one of your connections who stumbles upon the same site or is targeted by the same type of scam. At the very least, it could draw attention to the fact that scam sites affect real people. A post on Facebook about a close call you had with a scam might better equip your network to avoid any dangerous entanglements. If it does, they’ll thank you.

Continue Reading

Startups

LastPass hacked, OpenAI opens access to ChatGPT, and Kanye gets suspended from Twitter (again) • TechCrunch

Published

on

Aaaaand we’re back! With our Thanksgiving mini-hiatus behind us, it’s time for another edition of Week in Review — the newsletter where we quickly wrap up the most read TechCrunch stories from the past seven(ish) days. No matter how busy you are, it should give you a pretty good idea of what people were talking about in tech this week.

Want it in your inbox every Saturday morning? Sign up here.

most read

Instafest goes instaviral: You’ve probably been to a great music festival before. But have you been to one made just for you? Probably not. Instafest, a web app that went super viral this week, helps you daydream about what that festival might look like. Sign in with your Spotify credentials and it’ll generate a promo poster for a pretend festival based on your listening habits.

LastPass breached (again): “Password manager LastPass said it’s investigating a security incident after its systems were compromised for the second time this year,” writes Zack Whittaker. Investigations are still underway, which unfortunately means it’s not super clear what (and whose) data might’ve been accessed.

ChatGPT opens up: This week, OpenAI widely opened up access to ChatGPT, which lets you interact with their new language-generation AI through a simple chat-style interface. In other words, it lets you generate (sometimes scarily well-written) passages of text by chatting with a robot. Darrell used it to instantly write the Pokémon cheat sheet he’s always wanted.

AWS re:Invents: This week, Amazon Web Services hosted its annual re:Invent conference, where the company shows off what’s next for the cloud computing platform that powers a massive chunk of the internet. This year’s highlights? A low-code tool for serverless apps, a pledge to give AWS customers control over where in the world their data is stored (to help navigate increasingly complicated government policies), and a tool to run “city-sized simulations” in the cloud.

Twitter suspends Kanye (again): “Elon Musk has suspended Kanye West’s (aka Ye) Twitter account after the latter posted antisemitic tweets and violated the platform’s rules,” writes Ivan Mehta.

Spotify Wraps it up: Each year in December, Spotify ships “Wrapped” — an interactive feature that takes your Spotify listening data for the year and presents it in a super visual way. This year it’s got the straightforward stuff like how many minutes you streamed, but it’s also branching out with ideas like “listening personalities” — a Myers-Briggs-inspired system that puts each user into one of 16 camps, like “the Adventurer” or “the Replayer.”

DoorDash layoffs: I was hoping to go a week without a layoffs story cracking the list. Alas, DoorDash confirmed this week that it’s laying off 1,250 people, with CEO Tony Xu explaining that they hired too quickly during the pandemic.

Salesforce co-CEO steps down: “In one week last December, [Bret Taylor] was named board chair at Twitter and co-CEO at Salesforce,” writes Ron Miller. “One year later, he doesn’t have either job.” Taylor says he has “decided to return to [his] entrepreneurial roots.”

audio roundup

I expected things to be a little quiet in TC Podcast land last week because of the holiday, but we somehow still had great shows! Ron Miller and Rita Liao joined Darrell Etherington on The TechCrunch Podcast to talk about the departure of Salesforce’s co-CEO and China’s “great wall of porn”; Team Chain Reaction shared an interview with Nikil Viswanathan, CEO of web3 development platform Alchemy; and the ever-lovely Equity crew talked about everything from Sam Bankman-Fried’s wild interview at DealBook to why all three of the co-founders at financing startup Pipe stepped down simultaneously.

TechCrunch+

What lies behind the TC+ members-only paywall? Here’s what TC+ members were reading most this week:

Lessons for raising $10M without giving up a board seat: Reclaim.ai has raised $10 million over the last two years, all “without giving up a single board seat.” How? Reclaim.ai co-founder Henry Shapiro shares his insights.

Consultants are the new nontraditional VC: “Why are so many consultant-led venture capital funds launching now?” asks Rebecca Szkutak.

Fundraising in times of greater VC scrutiny: “Founders may be discouraged in this environment, but they need to remember that they have ‘currency,’ too,” writes DocSend co-founder and former CEO Russ Heddleston.

Source link

Continue Reading

Startups

Building global, scalable metaverse applications

Published

on

Previously we talked about the trillion-dollar infrastructure opportunity that comes with building the metaverse — and it is indeed very large. But what about the applications that will run on top of this new infrastructure?

Metaverse applications will be very different from the traditional web or mobile apps that we are used to today. For one, they will be much more immersive and interactive, blurring the lines between the virtual and physical worlds. And because of the distributed nature of the metaverse, they will also need to be able to scale globally — something that has never been done before at this level.

In this article, we will take a developer’s perspective and explore what it will take to build global, scalable metaverse applications.

As you are aware, the metaverse will work very differently from the web or mobile apps we have today. For one, it is distributed, meaning there is no central server that controls everything. This has a number of implications for developers:

Event

Intelligent Security Summit

Learn the critical role of AI & ML in cybersecurity and industry specific case studies on December 8. Register for your free pass today.


Register Now

  • They will need to be able to deal with data that is spread out across many different servers (or “nodes”) in a decentralized manner.
  • They will need to be able to deal with users that are also spread out across many different servers.
  • They will need to be able to deal with the fact that each user may have a different experience of the metaverse, based on their location and the devices they are using due to the fact not everyone has the same tech setup, and this plays a pivotal role in how the metaverse is experienced by each user.

These challenges are not insurmountable, but they do require a different way of thinking about application development. Let’s take a closer look at each one.

Data control and manipulation

In a traditional web or mobile app, all the data is stored on a central server. This makes it easy for developers to query and manipulate that data because everything is in one place.

In a distributed metaverse, however, data is spread out across many different servers. This means that developers will need to find new ways to query and manipulate data that is not centrally located.

One way to do this is through the blockchain itself. This distributed ledger, as you know, is spread out across many different servers and allows developers to query and manipulate data in a decentralized manner.

Another way to deal with the challenge of data is through what is known as “content delivery networks” (CDNs). These are networks of servers that are designed to deliver content to users in a fast and efficient manner.

CDNs are often used to deliver web content, but they can also be used to deliver metaverse content. This is because CDNs are designed to deal with large amounts of data that need to be delivered quickly and efficiently — something that is essential for metaverse applications.

Users and devices

Another challenge that developers will need to face is the fact that users and devices are also spread out across many different servers. This means that developers will need to find ways to deliver content to users in a way that is efficient and effective.

One way to do this is through the use of “mirrors.” Mirrors are copies of the content that are stored on different servers. When a user requests content, they are redirected to the nearest mirror, which helps to improve performance and reduce latency.

When a user’s device is not able to connect to the server that is hosting the content, another way to deliver content is through “proxies.” Proxies are servers that act on behalf of the user’s device and fetch the content from the server that is hosting it.

This can be done in a number of ways, but one common way is through the use of a “reverse proxy.” In this case, the proxy server is located between the user’s device and the server that is hosting the content. The proxy fetches the content from the server and then delivers it to the user’s device.

Location and devices

As we mentioned before, each user’s experience of the metaverse will be different based on their location and the devices they are using. This is because not everyone has the same tech setup, and this plays a pivotal role in how the metaverse is experienced by each user.

For example, someone who is using a virtual reality headset will have a completely different experience than someone who is just using a desktop computer. And someone who is located in Europe will have a different experience than someone who is located in Asia.

Though it may not be obvious why geographical location would play a part in something that is meant to be boundless, think of it this way. The internet is a physical infrastructure that is spread out across the world. And although the metaverse is not bound by the same physical limitations, it still relies on this infrastructure to function.

This means that developers will need to take into account the different geographical locations of their users and devices and design their applications accordingly. They will need to be able to deliver content quickly and efficiently to users all over the world, regardless of their location.

Different geographical locations also have different laws and regulations. This is something that developers will need to be aware of when designing applications for the metaverse. They will need to make sure that their applications are compliant with all applicable laws and regulations.

Application development

Now that we’ve looked at some of the challenges that developers will need to face, let’s take a look at how they can develop metaverse applications. Since the metaverse is virtual, the type of development that is required is different from traditional application development.

The first thing that developers will need to do is to create a “space”. A space is a virtual environment that is used to host applications.

Spaces are created using a variety of different tools, but the most popular tool currently is Unity, a game engine used to create 3D environments.

Once a space has been created, developers will need to populate it with content. This content can be anything from 3D models to audio files.

The next step is to publish the space. This means that the space will be made available to other users, who will be able to access the space through a variety of different devices, including desktop computers, laptops, tablets, and smartphones.

Finally, developers will need to promote their space. This means that they will need to market their space to users.

Getting applications to scale

Since web 3.0 is decentralized, scalability is usually the biggest challenge because traditional servers are almost impossible to use. IPFS is one solution that can help with this problem.

IPFS is a distributed file system used to store and share files. IPFS is similar to BitTorrent, but it is designed to be used for file storage rather than file sharing.

IPFS is a peer-to-peer system, which means that there is no central server. This makes IPFS very scalable because there is no single point of failure.

To use IPFS, developers will need to install it on their computer and add their space to the network. Then, other users will be able to access it.

The bottom line on building global, scalable metaverse applications

To finish off, the technology to build scalable metaverse applications already exists; but a lot of creativity is still required to make it all work together in a user-friendly way. The key is to keep the following concepts in mind:

  • The metaverse is global and decentralized
  • Users will access the metaverse through a variety of devices
  • Location and device management are important
  • Application development is different from traditional development
  • Scalability is a challenge, but IPFS can help

Clearly, we can’t have an article series about building the metaverse without discussing NFTs. In fact, these might be the key to making a global, decentralized, metaverse work. In our next article, we will explore how NFTs can be used in the metaverse.

By keeping these concepts in mind, developers will be able to create metaverse applications that are both user-friendly and scalable.

Daniel Saito is CEO and cofounder of StrongNode

Source link

Continue Reading

Trending

URGENT: CYBER SECURITY UPDATE