Chatbots are seemingly taking over the internet and mobile applications. First-time visitors to a site or app now are often greeted by a cheery chatbot seeking to answer questions or guide them through an onboarding process.
When site or app users have problems, retailers, banks, and other businesses are asking users if they want to chat rather than call. More often than not, this routes the users into an initial conversation with a chatbot.
Is sending a customer or user to a chatbot something they actually want?
A debate continues to rage about whether chatbots are viewed as annoying or helpful by customers. In fact, both views are true. And the determining factor for whether a user likes interacting with a chatbot — or even prefers it to a human interaction — depends entirely on the context.
Two Paths Diverged in the AI Chatbot Road: T-Mobile’s “No Robots” vs. BofA’s Erica Personal AI
On August 15, 2018, upstart mobile voice and data provider T-Mobile announced it would ban all robots and automated systems from direct interactions with customers on support calls and chats.
“There are no robots or automated phone menus. No getting bounced around from department to department. No shouting “representative,” crowed T-Mobile in a press release.
At the same time, T-Mobile expanded its live customer support hours to 24/7, going counter to trends limiting human support agent times to working hours. Since the move, T-Mobile claims it has achieved higher levels of customer satisfaction and improved customer retention/
Survey after survey has found people tolerate automated customer support through conversational AI and chatbots but don’t love it and don’t prefer it.
On the other extreme, some large enterprises have taken an alternative path of going all-in-on Conversational AI and chatbots. In June 2018, Bank of America launched Erica (a play on the last part of the word America).
The bot resides within Bank of America’s mobile banking app rolled out to all users. By December 2019, 10 million of the Bank of America’s mobile users had activated Erica within their apps and were interacting with the chatbot.
One of the more sophisticated chatbots on the market, Erica allows users to ask questions by voice by text messages, or simply to navigate through tap menus.
Surveys of users showed satisfaction with Erica of over 80%, which is a staggering figure in the often challenging world of financial services NPS scores.
So, which is the right way to deliver a superior Customer Experience in the most efficient manner? T-Mobile’s add-more-humans approach or Bank of America’s deep embrace of expensive but powerful AI?
Do customers prefer to work with bots or humans? The truth is complicated.
Surveys have indicated that customers prefer to speak with humans for support needs.
Conversely, by a variety of metrics, customers are growing more accustomed to chatbots and Conversational AI.
Over time, the reality is chatbots will handle a greater and greater portion of customer interactions and will also become an indispensable tool for human support and sales agents – virtually merging into one support continuum.
Equally important, in technology the true test of adoption is not what surveys say but what users do. Increasingly, that means talking or texting with chatbots rather than waiting to talk to a human,
Consumer Trends Driving Growth of Chatbots
More and more customers are choosing to interact with a chatbot over traditional phone support. This choice is being driven by broad technology user trends.
- Text More Popular Than Voice: We are more used to texting to communicate and even prefer it over voice calls. For large swathes of the populace, these modes are preferred over live conversations with humans. The rise of WhatsApp and Facebook Messenger, which have collectively over 3 billion users and are primarily used for chatting to avoid SMS fees, has further accelerated the trend.
- More Comfortable Interacting With Machines: We are growing more and more used to interacting with machines to ask questions or make requests. Asking Alexa to turn off the lights or telling Siri to call your mother is great. Asking an airline chatbot if any flights going into Denver are delayed has human-machine interactions that are now a normal part of our lives.
- Less Patience: We are becoming less patient. Netflix brings us millions of movie titles in a search window. We use an app to tell our Roomba to clean the house before we get home from work – and get our wish.Technology brings us instant gratification. As a result, we are less likely to be willing to spend time on hold or wait for a call back if a chatbot can answer our questions or take care of our problems. Related to this point is the fact that chatbots run 24/7 every day of the year. They scale up or down to meet demand and are always available. For consumers used to instant gratification, this is a powerful draw.
- Better Algorithms: The technology behind chatbots, Natural Language Processing (NLP), has progressed in leaps and bounds. This means we impatient humans can more easily interact with chatbots without having to repeat ourselves or resort to using multiple phrasings to get our request to register.New algorithms – most notably GPT3 – have come out that can reliably and economically be trained to understand specific types of subject matter and respond in ways that are remarkably close to normal human response.
According to the 2019 AI Index Report, published by a global consortium including Stanford University and Google, NLP can now comprehend passages of text better than humans. This also has allowed more advanced chatbots to handle complex, multi-step support tasks.
Chatbots can provide proactive guidance and even anticipate needs. A flight delay question might prompt an advanced chatbot to offer to book a hotel room for the night because it knew a late flight might mean a missed connection.
Three Simple Questions to Determine Context
Clearly, chatbot usage is growing and users are voting with their texts and their voices, indicating preference. That said, the desire for a human versus a bot remains highly context-dependent.
Context dictates what the chatbot is capable of doing in any given situation. Context is also variable and can shift with the state of the user as they run through a customer or support journey.
Understanding where a user is on the journey and their context can inform expectations of and proper usage of chatbots. Here are some simple questions to determine whether to use a chatbot or to what the limitations of a chatbot might be in a given situation.
Does the user want to talk to (or likely prefer to talk to) a representative?
This is a no-brainer. If they don’t want to chat, don’t make them chat. Ironically, many companies still push hard to drive users into chat support queues under the thesis that users will learn how to chat and adopt it (and save the company money).
Usually the act of saying “representative” is intentional enough that a company is far better off complying with their wishes. Here, too, AI can provide a guide.
Over time, companies can gather data about customer preferences and use that to better understand which conversational mode is best for what type of users based on any descriptive characteristics.
Can a chatbot recognize the user?
If a chatbot can recognize the identity of a user, then it can tap profile and historical data about the user to generate more bespoke solutions and conversation. Identifying the user is far easier when the user is on a mobile app or logged into a website or calling from a known phone number.
This question does limit advanced support to existing users rather than new users for which there is little history. But when it is possible to recognize the user and match them to a profile? If you can get a profile, then the canvas for chatbot to operate with is much more broad and interactions can be much more detailed rather than limited to simple keyword and menu-driven interactions.
Is the user asking a complicated or simple question?
Chatbots can quickly and easily dispatch more personalized answers to many simple questions. “When is my reservation?” or “What is my order status?” are easy to answer when the identity of a user is known and they are operating inside of a controlled environment.
In a similar manner, when the company is using a chatbot to replace a form or other structured information gathering exercise, then chatbots or Conversational AI can operate very effectively.
For more complex questions that involve multiple variables and may not be as easy to understand based on pure keyword analysis, more advanced chatbots that leverage NLP and conversational AI can increasingly provide back-and-forth support that is on par or better than human agents.
This is riding the curve of rapid improvements in AI, as demonstrated through the steady increase in the ability of AI systems to understanding complete even more complex natural language tasks as well or better than humans.
Conclusion: The Future of Customer Conversations Is A Hybrid Between Chatbots and Humans
T-Mobile may claim that it does not force anyone to talk to robots but in reality, its system can recognize automatically whether you are calling from your own device.
Behind the scenes, T-Mobile uses analytics and automation to help customer service agents do their jobs more quickly and efficiently. In this case, the chatbots may not be visible on the front but their output and enablement are visible on the back. Agents happen to act as the intermediary between the two.
This is the true future of chatbots – a technology that acts as a fluid interface somewhere in the customer journey to provide assistance. The recipient may be a customer talking to a chatbot or a support agent that has a chatbot automatically populating conversational snippets.
In this scenario, a company like T-Mobile can help agents work faster, answer questions more quickly, and breeze through the simplest queries. They can then save more time for the harder customers and questions that internal systems cannot automatically address.
BofA’s Erica can serve as a more forward presence, intercepting and deflecting simple inquiries. When a query grows too complex — and out of context for Erica — then the AI chatbot can easily route the request to a human support agent spending most of their time on tougher cases.
So which is right? Do customers prefer to talk to human agents rather than chatbots as the surveys indicate? Or do customers prefer to use chatbots to waiting to talk to humans, as the usage trends clearly indicate?
The answer is both. If customers are voting with their time spent and their immediate menu choices, they clearly do like intelligent chatbots more than waiting for a human. Meaning, they prefer chatbots and AI, given the right context and the right situation.
On the other hand, humans still and probably always will prefer live support agents when they have complex, nested and conditional questions to resolve.
These types of questions require the most advanced sort of conversational intelligence — one that even agents do better fulfilling when assisted by technology and AI behind the scene. The irony is that either way, customers are talking to chatbots — directly or indirectly.
The technology to improve both the experiences of T-Mobile and Bank of America customers is invariably the same under the covers. The sooner businesses realize that this is never an either-or-equation, the sooner they can determine where AI should sit in their Customer Experience stack.
How Blockchain Is Being Used With Smart Buildings – ReadWrite
Whether you realize it or not, many of us live in buildings with some smart capacity. You probably have at least one smart device in your home.
With the smart device industry set to grow by $65 billion by 2024, the odds are, you’ll add more of these devices. The true potential of smart homes lies in the ability of smart devices to communicate together — and that’s where blockchain technology comes in.
How Blockchain is Being Used With Smart Buildings
On the surface, smart technologies make individual tasks easier, but the potential is much larger than that. A smart device is effectively a sensor able to collect significant amounts of data about everything, from your energy use to how well-stocked your fridge is.
Smart Technology Works Better in Swarms
On its own, this data is valuable; when combined with data from other devices, its usability becomes game-changing. A properly connected smart home would be able to automatically adjust the heating to your preferences while minimizing bills, ordering your favorite groceries, monitoring and adjusting energy usage, sending repair notifications if something breaks, and much more.
Internet of Things (IoT) technologies are already used extensively in supply chain management. They help efficiently manage products passed through multiple stakeholders and verify that products are what the label says they are.
Catching Slave Labor in Fishing Supply Chains
One example where smart technology has been useful is in tracking fishing supply chains. The World Wildlife Federation (WWF) has used IoT to track sustainable tuna fishing.
The Western and Central Pacific tuna trade is rife with illegal fisheries — and, in some cases, slave labor — because tracking is either done via an easily-forged paper trail or not at all. However, savvy consumers and brands are demanding more accountability from the tuna industry.
The WWF’s branches in New Zealand, Australia, and Fiji have combined forces with blockchain software studio ConsenSys to implement secure traceability and track to address the problem.
Radio-frequency identification (RFID) or QR codes capture information as a fish moves through the supply chain from the boat to grocers. Tracking information is automatically saved in blockchain, making it nearly impossible to forge.
Privacy and Compatibility Remain a Concern
Although smart technology has many uses in enterprise settings, it becomes a thornier prospect for individuals. IoT devices collect huge amounts of data which can reveal a lot about their owners. Additionally, they are often poorly secured, creating significant security challenges.
Most smart devices must run on centralized platforms controlled by major tech companies, notably Amazon and Google.
There have been significant privacy concerns about both companies due to their access to an extraordinary amount of personal data.
Amazon Alexa’s Vulnerabilities
Setting aside concerns about microphones, Amazon’s voice-activated assistant Alexa also presents other significant security concerns.
Although Amazon provides some privacy protections, with 100 – 200 million Alexa devices and over 100,000 skills already deployed, there is a significant concern about malicious developers taking advantage of security holes.
For example, developer names aren’t verified, allowing a malicious developer to stage a phishing attack posing as a different company. This risk is especially high with some skills that link to email, banking, or social media accounts.
After a skill has been approved and added to the marketplace, a malicious developer can change its coding without getting Amazon’s approval or notifying the customer. Many developers also have misleading privacy policies — or none at all, meaning that customers will have no idea how their personally identifiable information will be used.
Lack of Device Compatibility
The second challenge is compatibility. Early adopters are painfully familiar with the concept of device divorce, where two smart devices cannot speak with another. Part of the problem is that Amazon and Google are used as primary smart home controllers, and there isn’t a platform-agnostic solution widely available to most consumers.
Blockchain Technology is the Missing Piece of the Puzzle
Blockchain technologies are working to provide the solution to these challenges and others since they can enable P2P connections without the need for a centralized validator.
With blockchain, it would be possible to connect numerous smart devices without being forced to hand that data directly over to the device manufacturer, mitigating privacy and security concerns. It can also provide increased transparency over how data is used, helping users understand what data their smart home is collecting and what it’s used for.
Blockchain technology is also hardware agnostic. Thus, it would be possible for users to pair together devices from different manufacturers without worrying about compatibility.
IOTA’s Tangle vs. Traditional Blockchain
One of the best examples of this vision is the IoT-focused blockchain IOTA.
It is important to understand that we are not talking about financial blockchain technology like Bitcoin. Blockchains based on traditional Proof of Work (PoW), like Bitcoin, lack the speed and scalability necessary to process the millions of data points produced by smart devices.
Instead, we are looking at smart device-focused technologies, most notably IOTA. IOTA uses a Tangle specifically designed for data and value transfer.
Blockchains like Bitcoin are essentially long chains of blocks containing transactions. The Tangle, on the other hand, is constructed as a directed acyclic graph (DAG), which is a collection of vertices connected by edges.
IOTA’s implementation is designed in such a way that each new transaction (vertice) must approve two previous transactions when it enters the Tangle. This eliminates the need for Proof of Stake (PoS) or PoW consensus methods.
Because these transactions don’t require always-online validators, they are feeless and contain metadata that makes them suitable for micropayments and data transfer.
IOTA is interesting because the technology is more mature than many other IoT-focused blockchain solutions. The project has experienced past problems, but the roll-out of its improved Tangle has allowed it to secure some important partnerships, primarily in areas designed to improve transparency.
Properly Validating Smart Device Data Is The First Step
IOTA’s most important partnership for smart homes is undoubtedly Project Alvarium. The biggest challenge posed by IoT — and smart devices in general — is the sheer volume of data collected. The vastness of information makes assessing what data is trustworthy and useful difficult, especially in an automated environment.
To solve this problem, Dell and IOTA teamed up to create Project Alvarium, designed to provide a simple way to assess the trustworthiness of data gathered.
Project Alvarium’s system logs every datapoint as it travels across the system. Each interaction is given a trust rating, which is logged on the IOTA Tangle to prevent tampering. This provides a simple way to find problems or deliberate tampering within a network of data.
Blockchain Can Help Resolve Security Concerns About Smart Security
When smart home users are certain that they can trust the data being generated by their devices, it opens up a world of opportunities that could transform our daily lives.
The most immediate use of blockchain technology is in improving building security. The most high-profile problem is undoubtedly Amazon’s Ring. In late 2020, dozens of people sued Amazon over accusations that their Ring doorbells had been breached.
The breach enabled hackers to watch people inside their homes and talk to individuals in the house over the Ring speakers.
The Blockchain Difference
Blockchain has been shown to resolve both the problem of data breaches as well as hacking takeovers. Capturing a blockchain-powered device would require compromising the entire blockchain itself compromised.
But proper validation, such as that proposed by IOTA, allows malicious devices to be pruned from the network, significantly improving security.
Additionally, blockchain could enable consumers to understand how their data is being used, helping to make smart devices more privacy-focused.
Smart Building Management Solutions are Already Being Tested
The value of blockchain technology becomes even bigger at scale. One of the most impactful uses of IoT and blockchain technology is in building management. Whether for an apartment building or an office building, it’s often difficult to effectively manage a building’s heating, lighting, and security in a way that minimizes waste.
Example: How Blockchain Could Manage Heating Bills
In a traditional setting, most buildings are managed centrally. If there is a unified heating system, it is often controlled by the local administration. Although this system is more efficient than individually-heated buildings, there is significant room for human error. That’s because the system is not optimized to account for more efficient heating higher up in the building as heat rises.
A network of heating sensors could be used to automatically measure the temperature in each apartment or office in a building. If the different thermostats could communicate with each other, it should be possible to input all the data into a blockchain solution.
A scheme like this would allow the building operators to create a proper heat map of the building and understand the most efficient usage of energy. It would also enable residents to access the data and understand why the system works the way it does.
Theoretically, it could also enable a user to select a target temperature for their apartment by leveraging rising heat from lower apartments.
Solutions on the Horizon
This kind of project is already being tested. For example, Brickschain offers several products that minimize difficulties with building management and handover on sale. There are also an increasing number of studies looking at how blockchain can be positively implemented into the building management process.
The Future of IoT: Many-to-Many Marketplaces
When buildings are utilizing IoT devices and blockchains, a bigger opportunity opens up: decentralized marketplaces.
Currently, it can be difficult to get the best deal on energy or heating bills because it is a marketplace with many customers but only a few providers. Switching providers can be difficult and doesn’t guarantee a competitive rate.
However, with blockchain, it would be possible to change providers based on real-time pricing data. This setup would create a competitive many-to-many environment where many providers are looking to sell energy to many customers. The competition among providers would drive down energy prices and improve overall efficiency in energy markets.
Swedish District Heating Study
Sweden has conducted studies to investigate the utility of blockchain for a district heating market. The setup allows apartment blocks already utilizing blockchain to automatically select the most affordable provider at any given moment, minimizing bills without requiring micromanagement.
The same concept could be applied to many aspects of building management.
One interesting idea is the concept of decentralized governance. This type of network could empower tenants and apartment owners to vote on changes to their apartment block’s management proceedings.
For example, renters could vote in favor of using only green energy sources or for changes to living space regulations. Building administrators could then better understand their occupants’ needs and create a better living environment for all involved.
Blockchain Will be Needed to do IoT Correctly
Adoption of IoT and smart technologies will likely increase. Governments like the UK are already pushing hard on smart meters and many of us have already adopted some form of smart technology in our homes.
This rush to adopt new technology will undoubtedly come with significant scaling problems as well as security concerns and significant privacy issues.
Additionally, a market dominated by a handful of major tech companies like Amazon and Google could prove damaging to the consumer in the long term.
To counter these eventualities, we’ll need a platform-agnostic solution that allows a more diverse field of producers to create new IoT devices.
Blockchain technology still represents the best way to utilize IoT for everyone’s benefit. If solutions like IOTA are implemented into existing smart homes, then we could build a new decentralized marketplace that will give us better control of our data, while improving the efficiency of our homes.
Image Credit: pixabay; thank you!
Social Sign-on: Sure, it’s convenient. But is it really safe? – ReadWrite
Remembering passwords is always a hassle, especially when you have innumerable websites that require logging in to view or interact with their content. To make the process simpler (as little as a couple of clicks), webmasters worldwide have accepted and implemented social logins on their websites.
Social Sign-on: Sure, it’s convenient. But is it really safe?
So, what exactly is social login? How different is it compared to the traditional method of inputting your credentials such as username, email address and password manually? More importantly, is it safe enough for use on all kinds of browsing activities?
In this article, we answer all the above questions and more, helping you understand what social sign-on is, and what the disadvantages of this convenient method are.
The history of social logins
Social sign-on as a method of hassle-free authentication has been around for over a decade now. Back in the nascent days of the modern internet in 2008, Facebook launched Facebook Connect, a service aimed at simplifying registrations on websites.
Once webmasters enabled FB Connect on their websites, visitors to the site would no longer need to fill up lengthy registration forms to sign up for the website’s offerings.
All they needed to do was connect their existing Facebook account to the website, enabling direct access to the site with a click of a button.
In 2009 and 2010, Twitter and LinkedIn respectively enabled their users to socially login to other sites using their existing social network credentials.
Google+ followed suit in 2011, and although no longer active as Google+, it still supports social sign-on using a Google account.
While it all sounds very convenient, social sign-on has many drawbacks and challenges that impact both website visitors and website owners.
Social Sign-on: The challenges and disadvantages
The Trust Factor
Most internet users do not trust the websites they browse to store and utilize their personal information safely and responsibly. Often, website visitors are concerned about how the information they have shared will be used.
In a June 2020 survey conducted by Insider Intelligence, 32% of US Facebook users felt that they somewhat disagreed that the platform could keep their data and privacy secure.
People tend to be wary of the private information they share online; they often resort to uploading falsified or inaccurate information about themselves on social media.
Considering that these social media sites do not verify or vouch for the authenticity of their user’s information, this could be less than ideal for a website looking for accurate data while accepting new user registrations.
In 2019, Facebook released data that said that 16% of the accounts on its platform are fake/duplicate accounts created by individuals or companies. What’s more worrisome are the findings of the research team at NATO StratCom that suggest 95% of the reported fake accounts still continued to remain active, with no action taken by the social media website.
With no checks on the actual profile that’s being used to socially sign-on to your website, you could soon have an imposter, Donald Trump or Joe Biden signing up for your global warming newsletter or purchasing a bag of your freshly powdered Mexican coffee.
Not everyone’s social — nor on social
While we talk about social media, we need to understand that although it is a global phenomenon with an insanely large number (read 3.6 billion) of people using it, there is still a sizeable chunk (>50%) of the population that is not on social media.
Using a restrictive method, you risk alienating a section of society that could be your potential target audience.
Transfer of Power
Enabling social sign-on seems pretty enticing at first, considering it would cut down your authentication work significantly. But this very ‘benefit’ could end up costing you dearly, as you lose control over your visitors’ data to a third-party service provider, i.e., the social media network.
Should there be any downtime at the social media service’s end, your website visitors would be stranded, unable to login to your site or access their data?
Access Control Issues
Many internet access places tend to have controls in place when it comes to accessing social media. For example, corporate and educational networks generally block access to social websites. Certain countries like Iran, China, Syria, and North Korea have blanket bans on the most popular social websites.
Social sign-on still depends on an API call-back to the social networking site to authenticate the user. Thus, by having social sign-on set up on your website, visitors authenticating on your site through these networks would end up facing a website with broken functionality.
Social media accounts are often the target of several hacking and phishing attempts. Thus, if your user’s social media account is hacked, it could lead to their account on your site being compromised as a result.
A University of Maryland study revealed a hacking attempt every 39 seconds on average, affecting a third of Americans every year.
Hacked social accounts could have an adverse impact on your website as well, by performing activities that might eat up your server resources or corrupt your files, if your security is not up to the mark. Secure authentication is the need of the hour, and knowledge of the security practices will help solve these concerns.
Too much to choose
People use many social media websites, so keeping a single social login can be counterproductive. However, providing multiple methods to login could likely confuse or overwhelm your visitor, leading to lower conversion or sign-up rates.
Lesser data to work with
Using a social sign-on for your website would mean limited access to user data, especially email. Not every social media network allows websites to access the customer’s email address. For businesses that rely on customer information for lead generation, this would be a major deal-breaker.
Awareness of all the security practices and malpractices (sawolabs dotcom) will help educate users as well as the website owners.
If not social sign-on, then what?
All the above drawbacks would make webmasters question the efficacy of social sign-on. But then, is there a better alternative that does not include such shortcomings?
Say hello to passwordless authentication powered by SAWO Labs. A new-age solution designed to address all concerns of security, compatibility and functionality.
Image Credit: yellow graphic — from author; thank you!
Top Image Credit: karolina grabowska; pexels; thank you!
3 Ways Companies Can Be More Sustainable – ReadWrite
I’m thinking about our planet today — I think about our planet every day. Our planet is hurting, and many businesses are encouraging their employees to live more sustainably. According to the Environmental Protection Agency (EPA), industry and agriculture account for approximately 32 percent of direct emissions.
3 Ways Companies Can Be More Sustainable
Here are a few ways companies can encourage sustainability in their employees and work to lower the remaining 68 percent.
Employ a hybrid work model
With so many people working from home due to the COVID-19 pandemic, we’ve inadvertently been doing Earth a huge favor. The EPA shows that transportation is responsible for 28 percent of greenhouse gas emissions, with about half of that coming from personal vehicles that burn gasoline and diesel. Because many companies instituted a work from home policy, there were fewer cars on the road and fewer greenhouse gas emissions. Companies can continue this progress by instituting a hybrid work model once the pandemic is finally over.
Reduce waste in the office
One big way to reduce waste in the office is by offering snack and drink options that eliminate single-use plastic. For example, TechnologyAdvice uses a Bevi machine in the office, offering still, sparkling, and flavored water without single-use plastic. You might also consider snacks that don’t need to be individually packaged, like fruits or nuts.
While you may not be able to completely eliminate office waste, you can work to offset the waste you do generate. Make it easy for employees to recycle and encourage them to do so. You can create an employee-led recycling program, keep an “I don’t know” bin for those items that don’t always fall into the normal categories, and create challenges around recycling goals.
Continual education about climate change
However you decide to encourage sustainability in your office, it’s important that both you and your employees engage in continual education about climate change. Thanks to the different forms of media available today, educating yourself about climate change has never been easier.
For podcast listeners, consider checking out How To Save A Planet. It’s a Spotify original podcast hosted by scientist Dr. Ayana Elizabeth Johnson and journalist Alex Blumberg, and it is the exact opposite of what people think when they hear “climate change resource:” it’s inspiring instead of depressing, entertaining, so accessible, and has great intro music.
Another Earth-friendly podcast you should listen to is Stories for Earth, which examines how climate change is discussed in pop culture.
If you like documentaries, check out Before the Flood, which was made by Leonardo Di Caprio and National Geographic. If you are a reader, consider these three: No One Is Too Small To Make A Difference by Greta Thunberg, All We Can Save edited by Ayana Elizabeth Johnson and Katherine Hayhoe, and The Future Earth by Eric Holthaus.
Image Credit: karolina grabowska; pexels; thank you!