Microsoft’s Bing AI just got a smart new feature, namely the ability to accept voice input on desktop PCs.
Voice input was already supported on mobiles, but now you can engage with the Bing chatbot asking queries via a microphone attached to your desktop PC. What’s more, the chatbot can reply with its own voice response (as Windows Central reports).
There are only a limited number of languages supported for voice input, mind you, at this stage. Those are as follows: English, French, German, Japanese and Mandarin.
We can doubtless expect more languages to be covered in the future, as Microsoft builds out this feature.
To use voice input, simply click the microphone icon in the Bing Chat box, and say your piece, then wait for the AI’s digitized voice response. The experience remains the same otherwise, as you might guess.
Analysis: Voice of the future
This is an important feature for the Bing AI, as it means that sessions with the chatbot feel like a more natural exchange – you’re talking, and the AI is replying with its voice. In other words, it’s more like a conversation than a search engine-like experience, which is Microsoft’s overall goal.
With the facility being on mobile already, it was only a matter of time before it was ported over for desktop users. More broadly, voice has been an area Microsoft has focused on considerably – witness the swift progress with Voice Access in Windows 11 in recent times (and in-line dictation in Microsoft Word, for that matter).
When Copilot lands in Windows 11 (supposedly testing will begin later in June), it’s easy to envisage that this infusion of AI will come with more voice-based controls, too.
However, the next big move for the Bing chatbot will be its introduction to other browsers besides Edge (with official support for Chrome and other major browsers inbound, as opposed to third-party plug-ins which are the only option currently).
Microsoft’s Bing AI is about to appear in all major web browsers according to a new report.
This comes from Windows Latest, which tells us that according to sources, the Bing chatbot will no longer be exclusive to Edge, but will be available in Chrome, Safari, and Firefox – all the main browsers – at some point this week (in a few days apparently).
Take that with a pinch of seasoning, naturally, but we already heard from Mikhail Parakhin, Microsoft’s head of Advertising and Web Services, last week, who informed us that “hopefully” the first experiments in enabling third-party browsers would be happening soon.
So, it seems that hope is now a reality, or is about to become one, with Windows Latest further reporting that Microsoft actually tested Bing AI in Apple’s Safari browser over the past weekend.
If you blinked, you’d have missed this, though, as the test was a brief one.
Windows Latest also received an email through, apparently sent to some Bing AI mobile users, which mentions new features inbound for the AI. That includes the idea of “characters with personalities in Bing AI”, meaning a more in-depth choice than the simple creative, precise, or balanced personalities that currently grace the chatbot.
Microsoft is also planning to lift some restrictions, we’re told, so that could mean longer chat sessions with Bing AI are on the way, perhaps.
Analysis: Sarcastic mode? Oh yes, that’s real likely, we’re sure…
Windows Latest actually got to use Bing AI in the Safari test, and reports that it’s much the same experience as using the chatbot in Microsoft’s Edge browser. That’s pretty much what we’d expect, of course – there’s no reason it would be meaningfully different.
As we’ve discussed previously, it makes more sense for Microsoft to focus on driving usage of the Bing chatbot, than it does to use the bot as a lure to get people to switch to the Edge browser.
Yes, Edge is doubtless very important to Microsoft, but having its AI outgun Google’s Bard is surely a far more important consideration. And so having Bing AI in all the big browsers will help to that end, though we weren’t expecting this to happen quite as soon as this week. That would clearly indicate this is a real priority for Microsoft.
As for the idea of more varied personalities, this was something hinted at before in the very early days of Bing AI. Windows Latest points to leaked personalities that include ‘friendly’ and ‘sarcastic’ modes, though in the latter case, we’re skeptical as to whether this might be in the works.
Granted, sarcastic mode would be entertaining, certainly. But when folks have tried to get entertainment out of Bing AI in the past, pushing its buttons and boundaries, Microsoft has done its best to limit the chatbot’s more off-the-wall responses, and we’re not sure we see that changing anytime soon.
Don’t look now, but Microsoft is exploring the idea of pushing adverts with its Bing AI chatbot.
The ChatGPT-powered AI hasn’t been around for long, but has already achieved a measure of success driving more traffic to the Bing search engine, which is one of Microsoft’s main aims with the chatbot, naturally.
However, another target Microsoft is now sighting up is the addition of advertising to monetize the chatbot in a more concrete way, as evidenced by a freshly written post on the Bing blog site.
The content makes for worrying reading – though it’s a predictable avenue for Microsoft to assess, and we should note that this is very much early days still. The blog post talks about “exploring” the use of ads and that it’s mulling over “some early ideas,” so we shouldn’t get too carried away at this point.
The ideas Microsoft is toying with include making it so that if the user hovers over a link, it’ll pop up a panel containing further links to a publisher’s content. And also placing a “rich caption of Microsoft Start licensed content” next to a chat answer, meaning driving traffic to these Microsoft Start partners (and sharing ad revenue with that partner).
Microsoft further states: “We’re also exploring placing ads in the chat experience to share the ad revenue with partners whose content contributed to the chat response.”
Analysis: A case of greed before need?
Oh dear, oh dear. And another oh dear for good measure. Microsoft seems to be losing sight of the purpose of its ChatGPT-fueled AI here. It’s supposed to be a useful tool, a smart addition to Bing to let people search in a new way, and do a whole lot more to help them besides, with a bunch of other tricks up its sleeve (from knocking up artwork or poetry, to providing swift aid to gamers, kind of).
So, Bing AI is all about helping folks, right? Well, nobody believes Microsoft is doing this out of the goodness of its own heart. Of course there’s an agenda, and that’s to push Bing search to compete better with the highly dominant Google. Fair enough: Bing needs some kind of secret weapon, you can’t argue with that, and finally, it looks like the AI chatbot might be the answer.
Google is even worried, having rushed its rival AI, Bard, onto the stage, fumbling its lines in the process (and not getting nearly the same amount of attention as Bing AI, which is already well underway with updates being applied regularly by Microsoft).
But a measure of early success appears to have left Microsoft with dollar signs flashing before its eyes, perhaps blinding it to Bing’s original purpose. So now, we have Microsoft thinking about getting greedy (there’s evidence of this with the company’s attitude elsewhere regarding the AI, too). Bing traffic isn’t enough in the way of monetization, perhaps, so why not do adverts, too?
We’ll tell you why not – because people are tired of you trying to jam adverts into everything, Microsoft. Like your online services, or the Windows 11 interface, over and over, with the latest example coming earlier this week in the form of (veiled) ads for the Start menu. And now the Bing chatbot?
We’re not surprised, really, but we are getting fed up with the condition Microsoft seems to suffer from, the main symptom of which is a sickening and relentless compulsion to cram in adverts with its products and services. Let’s call it ‘ad nauseum’, and let’s further hope a cure can be found.
Clyde, Discord's cartoon mascot, is getting an AI makeover. On Thursday, the popular chat platform unveiled a wide collection of new AI tools and experiments including an OpenAI-powered chatbot update.
You may not use Discord every day, or at all, but it's become a go-to destination for people looking to have group discussions about topics like gaming, pets, anime, coding, and other nerdy pursuits. According to Discord, 150 million people chat on the service each month.
It's also become a key destination for AI content creation where, according to Discord, 30M users have experienced AI apps on the platform each month.
While most people are now familiar with OpenAI's DALL-E 2 AI image generation tool, Discord users have been creating images with Midjourney's generative AI image tools for just as long. The difference in Discord, though, is the community enhancement that revolves around AI content creation.
As Discord CEO Jason Citron explained it on Tuesday, Discord provides, “AI at your fingertips, together.”
Next week, Discord rolls out the OpenAI-enhanced Clyde as a free public experiment.
Discord Clyde AI (Image credit: Discord)
Like Bing's new Chatbot, Clyde utilizes ChatGPT's large language model to engage in conversation. You invoke Clyde, opt-in to use him on your server and in your group, and can then start asking Clyde questions much in the same way as you'd converse with ChatGPT.
The difference is that because Clyde operates in Discord's chat environment, it has to be aware of group discussion dynamics. Clyde will not just pipe in and interrupt a group chat but a mention of it will allow Clyde to join in. Clyde operates like a real Discord member and can in its responses include GIFs and emojis (ChatGPT in Bing can use emojis, but not GIFs).
Clyde can look up information to settle a dispute or assist with the group project. As with everything else on Discord, admins can easily disable Clyde.
While we've seen a lot of AI chatbot abuse, Discord's Citron prefaced the announcements by reiterating how they work hard to enable all of this “in a safe and trustworthy environment.”
Other AI experiments
Beyond chat, Discord is actively spreading AI capabilities throughout the platform, with many of them launching as limited experiments next week.
To help server moderators, Discord is enhancing its AutoMod tool which already uses preset keywords to actively ban content from group chats with AI that takes proactive moderation a step further.
Discord AutoMod AI (Image credit: Discord)
If, for example, you posted in your group chat that the group does not allow self-promotion or topics outside the core one, which might be sailboats, AutoMod AI could use that post to ban posts skirting those rules. It can even do it if the post is in a different language.
(Image credit: Discord)
Discord chats are often quite active and since no one can spend 24/7 monitoring them, Discord is adding a Conversations Summary Experiment. As the name suggests, it can with a prompt read through previous Discord discussions to surface who was chatting, and the context of all the messages.
If, for example, someone at the end of a long chat asks you if you're coming to “the event on Saturday,” you can ask the AI to summarize the previous chat activity to figure out what event they were talking about – and if you want to attend.
More AI tools
Discord Avatar Remix AI (Image credit: Discord)
We also got a sneak peek at a couple of interesting AI tools, one productivity-focused and the other mostly pure whimsy.
There's the Avatar Remix App, which will let you use a prompt to apply a generative image update to a group member's avatar. Discord showed us how it could easily add a birthday hat to one avatar image and a mustache to the other. The effect is realistic and pretty clever. Obviously, there might be some concern about image abuse here, but at least the tool is only usable on people who already opted into your group.
Discord plans to place the open-source code on GitHub so developers can fork, mix, and extend the Remix code.
(Image credit: Discord)
There's also a powerful-looking new integrated Whiteboard with AI. It'll be the first shared, real-time whiteboard space inside of Discord. The AI integration lets you use simple sketches and text prompts to generate rich, expressive images. Discord claims it can “solve the blank canvas syndrome” many people have with these AI image-generation tools.
To try out any of these AIs, though, you need to be on Discord. That means setting up an account, joining a server, and finding or setting up your own groups. It's not hard but for the uninitiated, Discord can be overwhelming. However, these cool tools might inspire you to give it a try.
Microsoft has deployed a new version of its Bing chatbot (v96) featuring improvements to make the AI smarter in a couple of key areas – and a big change has been flagged up as imminent, too.
Mikhail Parakhin, who heads up the Advertising and Web Services division at Microsoft, shared this info on Twitter (via MS Power User).
OK, it took longer than we initially expected, but finally Bing Chat v96 is fully in production. Give it a try! Now, onto fully shipping the tri-toggle…February 28, 2023
See more
So, what’s new with v96? Parakhin explains that users of the ChatGPT-powered Bing will now experience a ‘significant’ reduction in the number of times that the AI simply refuses to reply to a query.
There will also be “reduced instances of hallucination in answers” apparently, which is industry lingo meaning that the chatbot will produce fewer mistakes and inaccuracies when responding to users. In short, we should see less misinformation being imparted by the chatbot, and there have been some worrying instances of that occuring recently.
The other major news Parakhin delivers is that the so-called tri-toggle, known more formally as the Bing Chat Mode selector – featuring three settings to switch between different personalities of the AI Bing – is set to go live in the “next couple of days” we’re told.
Analysis: Long and winding road ahead
The ability to switch between a trio of personalities is the big change for the Bing chatbot, and to hear that it’s imminent is exciting stuff for those who have been engaging with the AI thus far.
As detailed previously, the trio of personalities available are labeled as Precise, Balanced, and Creative. The latter is set to provide a chattier experience, and Precise will offer a shorter, more typical ‘search result’ delivery, with Balanced being a middle road between the two. So, if you don’t like how the AI is responding to you, at least there will be choices to alter its behavior.
Various different versions of the Chat Mode selector have been tested, as you would imagine, and the final model has just been picked. This is now undergoing honing before release which should happen later this week as noted, but we’re guessing there’ll be plenty of further fine-tuning to be done post-release.
Certainly if the overall Bing AI experience has been anything to go by, as the whole project is, of course, still in its early stages, and Microsoft is chopping and changing things – sometimes in huge ways – seemingly without much caution.
The current tuning for v96 to ensure Bing doesn’t get confused and simply not reply will help make the AI a more pleasant virtual entity to interact with, and the same will hopefully be true for the ability to switch personalities.
At the very least, the Creative personality should inject some much-needed character back into the chatbot, which is what many folks want – because if the AI behaves pretty much like a search engine, then the project seems a bit dry and frankly in danger of being judged as pointless. After all, the entire drive of this initiative is to make Bing something different rather than just a traditional search experience.
It’s going to be a long road of tweaking for the Bing AI no doubt, and the next step after the personalities go live will likely be to lift that chat limit (which was imposed shortly after launch) to something a bit higher to allow for more prolonged conversations. If not the full-on rambles initially witnessed, the ones that got the chatbot into hot water for the oddities it produced…
ChatGPT has quickly become one of the most significant tech launches since the original Apple iPhone in 2007. The chatbot is now the fastest-growing consumer app in history, hitting 100 million users in only two months – but it's also a rapidly-changing AI shapeshifter, which can make it confusing and overwhelming.
That's why we've put together this regularly-updated explainer to answer all your burning ChatGPT questions. What exactly can you use it for? What does ChatGPT stand for? And when will it move to the next-gen GPT-4 model? We've answered all of these questions and more below. And no, ChatGPT wasn't willing to comment on all of them either.
In this guide, we'll mainly be covering OpenAI's own ChatGPT model, launched in November 2022. Since then, ChatGPT has sparked an AI arms race, with Microsoft using a form of the chatbot in its new Bing search engine and Microsoft Edge browser. Google has also quickly responded by announcing a chatbot, tentatively described as an “experimental conversational AI service”, called Google Bard.
These will be just the start of the ChatGPT rivals and offshoots, as OpenAI is offering an API (or application programming interface) for developers to build its skills into other programs. In fact, Snapchat has recently announced a chatbot 'called My AI' that runs on the latest version of OpenAI's tech.
For now, though, here are all of the ChatGPT basics explained – along with our thoughts on where the AI chatbot is heading in the near future.
What is ChatGPT?
ChatGPT is an AI chatbot that's built on a family of large language models (LLMs) that are collectively called GPT-3. These models can understand and generate human-like answers to text prompts, because they've been trained on huge amounts of data.
For example, ChatGPT's most recent GPT-3.5 model was trained on 570GB of text data from the internet, which OpenAI says included books, articles, websites, and even social media. Because it's been trained on hundreds of billions of words, ChatGPT can create responses that make it seem like, in its own words, “a friendly and intelligent robot”.
(Image credit: ChatGPT)
This ability to produce human-like, and frequently accurate, responses to a vast range of questions is why ChatGPT became the fastest-growing app of all time, reaching 100 million users in only two months. The fact that it can also generate essays, articles, and poetry has only added to its appeal (and controversy, in areas like education).
But early users have also revealed some of ChatGPT's limitations. OpenAI says that its responses “may be inaccurate, untruthful, and otherwise misleading at times”. OpenAI CEO Sam Altman also admitted in December 2022 that the AI chatbot is “incredibly limited” and that “it's a mistake to be relying on it for anything important right now”. But the world is currently having a ball exploring ChatGPT and, despite the arrival of a paid ChatGPT Plus version, you can still use it for free.
What does ChatGPT stand for?
ChatGPT stands for “Chat Generative Pre-trained Transformer”. Let's take a look at each of those words in turn.
The 'chat' naturally refers to the chatbot front-end that OpenAI has built for its GPT language model. The second and third words show that this model was created using 'generative pre-training', which means it's been trained on huge amounts of text data to predict the next word in a given sequence.
An illustration from Google’s 2017 research paper for the Transformer architecture, which ChatGPT is based on. (Image credit: Google)
Lastly, there's the 'transformer' architecture, the type of neural network ChatGPT is based on. Interestingly, this transformer architecture was actually developed by Google researchers in 2017 and is particularly well-suited to natural language processing tasks, like answering questions or generating text.
ChatGPT was released as a “research preview” on November 30, 2022. A blog post casually introduced the AI chatbot to the world, with OpenAI stating that “we’ve trained a model called ChatGPT which interacts in a conversational way”.
The interface was, as it is now, a simple text box that allowed users to answer follow-up questions. OpenAI said that the dialogue format, which you can now see in the new Bing search engine, allows ChatGPT to “admit its mistakes, challenge incorrect premises, and reject inappropriate requests”.
(Image credit: OpenAI)
ChatGPT is based on a language model from the GPT-3.5 series, which OpenAI says finished its training in early 2022. But OpenAI did also previously release earlier GPT models in limited form – its GPT-2 language model, for example, was announced in February 2019, but the company said it wouldn't release the fully-trained model “due to our concerns about malicious applications of the technology”.
OpenAI also released a larger and more capable model, called GPT-3, in June 2020. But it was the full arrival of ChatGPT in November 2022 that saw the technology burst into the mainstream.
How much does ChatGPT cost?
ChatGPT is still available to use for free, but now also has a paid tier. After growing rumors of a ChatGPT Professional tier, OpenAI said in February that it was introducing a “pilot subscription plan” called ChatGPT Plus in the US. A week later, it made the subscription tier available to the rest of the world.
ChatGPT Plus costs $ 20 p/month (around £17 / AU$ 30) and brings a few benefits over the free tier. It promises to give you full access to ChatGPT even during peak times, which is when you'll otherwise frequently see “ChatGPT is at capacity right now” messages during down times.
(Image credit: OpenAI)
OpenAI says the ChatGPT Plus subscribers also get “faster response times”, which means you should get answers around three times quicker than the free version (although this is no slouch). And the final benefit is “priority access to new features and improvements”, like the experimental 'Turbo' mode that boosts response times even further.
It isn't clear how long OpenAI will keep its free ChatGPT tier, but the current signs are promising. The company says “we love our free users and will continue to offer free access to ChatGPT”. Right now, the subscription is apparently helping to support free access to ChatGPT. Whether that's something that continues long-term is another matter.
How does ChatGPT work?
ChatGPT has been created with one main objective – to predict the next word in a sentence, based on what's typically happened in the gigabytes of text data that it's been trained on.
Once you give ChatGPT a question or prompt, it passes through the AI model and the chatbot produces a response based on the information you've given and how that fits into its vast amount of training data. It's during this training that ChatGPT has learned what word, or sequence of words, typically follows the last one in a given context.
For a long deep dive into this process, we recommend setting aside a few hours to read this blog post from Stephen Wolfram (creator of the Wolfram Alpha search engine), which goes under the bonnet of 'large language models' like ChatGPT to take a peek at their inner workings.
But the short answer? ChatGPT works thanks to a combination of deep learning algorithms, a dash of natural language processing, and a generous dollop of generative pre-training, which all combine to help it produce disarmingly human-like responses to text questions. Even if all it's ultimately been trained to do is fill in the next word, based on its experience of being the world's most voracious reader.
What can you use ChatGPT for?
ChatGPT has been trained on a vast amount of text covering a huge range of subjects, so its possibilities are nearly endless. But in its early days, users have discovered several particularly useful ways to use the AI helper.
Other language-based tasks that ChatGPT enjoys are translations, helping you learn new languages (watch out, Duolingo), generating job descriptions, and creating meal plans. Just tell it the ingredients you have and the number of people you need to serve, and it'll rustle up some impressive ideas.
But ChatGPT is also equally talented at coding and productivity tasks. For the former, its ability to create code from natural speech makes it a powerful ally for both new and experienced coders who either aren't familiar with a particular language or want to troubleshoot existing code. Unfortunately, there is also the potential for it to be misused to create malicious emails and malware.
We're also particularly looking forward to seeing it integrated with some of our favorite cloud software and the best productivity tools. There are several ways that ChatGPT could transform Microsoft Office, and someone has already made a nifty ChatGPT plug-in for Google Slides. Microsoft has also announced that the AI tech will be baked into Skype, where it'll be able to produce meeting summaries or make suggestions based on questions that pop up in your group chat.
Does ChatGPT have an app?
ChatGPT doesn't currently have an official app, but that doesn't mean that you can't use the AI tech on your smartphone. Microsoft released new Bing and Edge apps for Android and iOS that give you access to their new ChatGPT-powered modes – and they even support voice search.
The AI helper has landed on social media, too. Snapchat announced a new ChatGPT sidekick called 'My AI', which is designed to help you with everything from designing dinner recipes to writing haikus. It's based on OpenAI's latest GPT-3.5 model and is an “experimental feature” that's currently restricted to Snapchat Plus subscribers (which costs $ 3.99 / £3.99 / AU$ 5.99 a month).
OpenAI's CEO Sam Altman has confirmed that it's working on a successor to the GPT-3.5 language model used to create ChatGPT, and according to the New York Times this is GPT-4.
Despite the huge number of rumors swirling around GPT-4, there is very little confirmed information describing its potential powers or release date. Some early rumors suggested GPT-4 might even arrive in the first few months of 2023, but more recent quotes from Sam Altman suggest that could be optimistic.
For example, in an interview with StrictlyVC in February the OpenAI CEO said in response to a question about GPT-4 that “in general we are going to release technology much more slowly than people would like”.
He also added that “people are begging to be disappointed and they will be. The hype is just like… We don’t have an actual AGI and that’s sort of what’s expected of us.” That said, rumors from the likes of the New York Times have suggested that Microsoft's new Bing search engine is actually based on a version of GPT-4.
While GPT-4 is unlikely to bring anything as drastic as graphics or visuals to the text-only chatbot, it is expected to improve on the current ChatGPT's already impressive skills in areas like coding. We'll update this article as soon as we hear any more official news on the next-gen ChatGPT technology.
1 844.742.7448 
Email Us
Remote Support 
Contact
