Will OpenAI’s “gpt-realtime” set a new benchmark for AI Voice?

OpenAI has introduced gpt-realtime, a new cutting-edge speech-to-speech model, alongside the general availability of its Realtime API. This release marks a significant step forward in the evolution of voice AI, particularly for enterprise applications such as customer support and conversational agents. They announced this in a video broadcast which you watch below.

SIP Telephony Support: Lowering the Barrier to Entry

One of the most notable updates they annunced was the addition of SIP telephony support, which aims to simplify the process of building voice-over-phone applications. Developers will be able to integrate phone numbers directly into OpenAI’s SIP interface, streamlining deployment and reducing the need for complex telephony infrastructure. As it develops, this could reshape the competitive landscape, especially for startups that previously relied on expensive and bespoke integrations to differentiate their offerings.

A Unified Model for Natural Interaction

Gpt-realtime apart will feature an end-to-end architecture which will set it apart to how such integrations work today. Unlike traditional systems that chain together speech recognition, language processing, and text-to-speech, OpenAI’s new model will handle everything in a single pass. This will result in much faster response times, more natural audio, and improved emotional nuance (one of the biggest limitations today) meaning it will be capable of interpreting laughter, stress, worry, pauses, and tone shifts.

Open AI so it will also be highly configurable. Developers will be able to adjust pacing, tone, and persona, enabling more tailored and brand-consistent voice experiences.

Considerations for Enterprise Adoption

While the capabilities are lok super impressive, these models will still be expensive to start with anyway. Pricing is expected to be $32 per million input tokens and $64 per million output tokens which is significantly higher than traditional chained models. Additionally, the unified architecture offers less modularity and observability, which may limit flexibility for teams that require fine-grained control over model behavior or voice switching.

In a blog post from CX Today, Alex Levin, CEO at Regal is quoted saying,the cost of the speech-to-speech model is still approximately four times higher than chaining a speech-to-text (STT), large language model (LLM), text-to-speech (TTS) pipeline for Voice AI Agents

Strategic Implications

OpenAI’s latest release is a clear signal of intent: to make voice AI more accessible, performant, and enterprise-ready. Given Mcirosoft and other leading Cloud giants, close relationships with Open AI, we can also expect them to eventally add support for such models meaning customers that leverage, for example Microsoft 365 Copilot and Azure AI will likely gain support for this in the near future too through tools like Microsoft Dynamnics and Copilot Studio.

For organisations exploring and wanting to experiment more with conversational based automation, gpt-realtime promises to offers a powerful new toolset whilst talking the technolgy closer to human voice.

As always, the key lies in aligning technology choices with business goals, recognising ROI and customer expectation and keeping ahead of the curve as the landscape evolves and the pace of AI maturity and adoption contines to accelerate.


Sources: 
CX Today – OpenAI’s Latest Moves Put Many Voice AI Startups on Notice
Open AI – YouTube Video:
Open AI Blog

Microsoft makes OpenAI o1 model free for Copilot users.

OpenAI’s most advanced AI model “o1” which is known for its problem solving and deeper thinking has been available behind a $20 per month ChatGPT premium subscription. ChatGPT premium gives limited acess for $20 a month and unlimited access for $200 a month.

Copilot let’s you use it for free.

Microsoft has a tight partnership with OpenAI and is also on a mission to put their AI (Copilot) across every Microsoft Service it offers with huge capability and features even on theor “free” tiers.

Copilot Consumer Pro users have had access to Think Deeper (which uses the o1 model) for the past 12 months, but Microsoft have now made this feature free to everyone including those using the free version of Copilot.

To access it, you need to simply head ovee to Copilot on the web, (or via the mobile app) and ensure you are signed in with a Microsoft account (MSA). You then get completed free access to the Think Deeper search (which uses the o1 model).

How to get Microsoft Copilot

To get Copilot, head to the web (you actually find Copilot in the Edge browser) and go to https://copilot.microsoft.com or head over to you phones app store and search for Copilot and install it.

You need to be signed in with your Microsoft account to use these features.

Using o1 features aka Think Deeper

Once in Copilot, use the AI chat as you would before (or like you did in ChatGPT) and you will see a “think deeper” button inside the text input box.

Using Copilot’s Think Deeper (ChatGPT model o1)

Selecting it activates the o1 reasoning model. As it processed your prompt, you also get a spinning symbol since searches and responses using o1 are more thorough that with GPT 4 and typpically take around 30 secs.

Using Copilot’s Think Deeper.

This is Microsoft’s way of letting you know that you’re in for around a 20-30 seconds wait. If you don’t need deep search so for normal use), toggle this back off to go back to the super fast GPT-4o version…

So what can o1 do then?

The deep thinker feature of Microsoft Copilot is much better for more complex tasks and research due to the o1 model ability for in depth reasoning. 

As such it is simply better for solving complex issues like math, logic or science, for analysing or creating long or richer documents and reports or for code creation and debug. The best way to test this is to run two Copilot Windows side by side and test out the same prompt with and without Think Deeper enabled.

Content created with o1 is also more “accurate” with far less AI hallucinations (aka, making things up).

Why do many GPTs Hallucinate? In general, GPT models learn by mimicking patterns in their training data (huge amounts of data). The o1 model uses a different technique called reinforcement learning, whereby it's language model works things out (though it's training) by rewarding the right answers and penalising wrong ones. This takes longer through the iterative and testing process. Once done the model  moves through queries in a step-by-step fashion much like human problem  solving. 

o1 limitations?

It is worth noting that o1 isn’t quite on the same level as ChatGPT in some areas. It is less effective with factual knowledge and is currently less able to search the Internet and cannot process files and images.

What about DeepSeek?

The big story this week has of course been DeepSeek, a controversial Chinese AI firm that has announced and launched their own GPT-4 and o1 rivals that have been supposedly built at a fraction of the cost of OpenAI, Google and other US models, shaking share prices, disrupting the market and rasing many questions.

What is more is more is that DeepSeek models are claimed to be more advanced and faster than GPT-4o and smarter that o1.

The advent of DeepSeek has sent shockwaves through the tech industry. Global stock markets have reeled, sparking a cascade of investigations and looming threats of bans.

Yet, the bot hasn’t been without its champions. Interestly, Microsoft – OpenAI’s top financial invester and partner  – has already embraced the DeepSeek R1 reasoning model, and has integrating it into Azure AI Foundry and also GitHub.

These platforms, beloved by developers for fostering advanced AI projects, now stand as the new playground for DeepSeek’s innovative potential.

DeepSeek logo

Open AI Strikes Back

In the wake of its free mobile app’s viral triumph, OpenAI’s CEO Sam Altman swiftly revealed plans to accelerate the rollout of new releases to keep ahead of its new Chinese competitor.

OpenAI are not standing still either though. Et the end of December 2024, month, they began  trialing twin AI models, o3 and o3 mini. Remarkably, the former has surpassed o1 in coding, mathematics, and scientific capabilities, marking a significant advancement in their AI prowess.

There is no doubt this is an area that doesn’t stand still. By the time I click publish this post will likely already be out of date!


DeepSeek has certainly ignited an even greater sense of urgency within the already dynamic AI sector which moves and evolves on an almost daily basis.

OpenAI launches ChatGPT Gov for U.S. Government Agencies

This week, in a significant development amidst the backdrop of intensifying AI competitiveness, OpenAI has unveiled ChatGPT Gov

This customised version of the AI-powered chatbot platform is tailored specifically for U.S government agencies, providing them with an advanced tool to access and utilise Open AI technology.

Microsoft announced back in 2023 that its Azure OpenAI Service was available for Azure Government customers. The new service will allow government agencies to use generative AI capabilities in a way that meets security and privacy requirements. OpenAI can now also be deployed in Microsoft Gov Data Centres.

What is ChatGPT Gov?

So ChatGPT Gov mirrors many capabilities of OpenAI’s enterprise-focused tier, ChatGPT Enterprise. By leveraging this platform, Open AI say that government agencies can deploy specific OpenAI models on both Microsoft Azure commercial and government clouds as well as using Microsoft’s own Azure AI models for example. This integration brings enhanced management of security, privacy, and compliance concerns, which is crucial for handling non-public sensitive and classified data.

ChatGPT Gov also aims to streamline internal authorisation processes, making it easier for agencies to implement OpenAI’s tools effectively as easily with the relevant guard rails in place.

What about Azure Open AI for Gov

Microsoft Open AI, enables federal, state, and local government agencies to use GPT-3, GPT-4 and 4o along with embeddings via the Azure OpenAI Service REST APIs. This capability helps to improve natural language-to-code translation, semantic search, content generation, and summarisation and for Gov to build and use Microsoft Open AI services across Gov cloud.

Gen AI in Government

Since its introduction, ChatGPT has already seen extensive adoption across the U.S. government as well as here in the UK. I’m. Personally working with a dozen or so local governments and councils here in the UK on AI adoption.

Open AI says that more than 90,000 users from more than 3,500 federal, state, and local agencies have collectively sent over 18 million messages to support their daily operations. This widespread usage demonstrates Open AI and Microsoft’s potential to transform government workflows and decision-making processes.

What about Copilot

While ChatGPT Gov offers a robust AI solution for government agencies, it’s worth exploring how Microsoft 365 Copilot also serves these needs. Microsoft

For many organisations using or exploring Gen AI tools like ChatGPT, many are using a combination of tools and services from different vendors. Open AI and Microsoft are tightly partnered.

Microsoft 365 Copilot is built on Open AI (which in turns runs in Microsoft Azure) and integrates seamlessly with existing Microsoft 365 tools, providing personalised assistance across a range of applications such as Word, Excel, and Outlook and also supports the building (both professional and low code) of autonomous AI agents, and scheduled prompts (coming soon).

This integration ensures that users can enhance productivity and streamline tasks within the familiar Microsoft ecosystem.  So how does Microsoft 365 Copilot differ to ChatGPT?

ChatGPT

  • Targeted for AI-powered chat and conversation but also supports connectors and extebsibikity to other services via extensions and APIs.
  • Recently launched the ChatGPT Gov version for U.S. government agencies and are expected to do similar in other global regions.
  • Deployable on Microsoft Azure commercial and government clouds
  • Doesn’t provide native integration into line of business office apps and services like Office 365, Power Platform and Fabric.

Microsoft 365 Copilot

  • Customised version of ChatGPT that runs in Microsoft 365 Tennant boundaries.
  • Provides chat based conversations and access to company agents and connectors on PAYG basis or via Microsoft 365 Copilot subscription.
  • Embedded within Microsoft 365 applications like Word, Excel, Teams, and Outlook as well as Dynamics 365 and Power Platform.
  • Designed to enhances productivity and efficiency within the existing Microsoft ecosystem and seen as add on to Microsoft 365 on a per user per month billing method.
  • Can provides contextual assistance and automation for daily tasks and workflows through agents and autonomous agents (public preview).

In conclusion, while both ChatGPT and Microsoft 365 Copilot are powerful AI tools, they cater to slightly different use cases

US Government agencies may find ChatGPT Gov particularly beneficial for secure, AI-driven interactions (in place of the general version of ChatGPT) , whereas Microsoft 365 Copilot excels in enhancing productivity and providing natively and seemlessly integration into their wider app services and data. Gov agencies using Microsoft 365 Copilot and Azure AI or Open AI deployed in Azure also benefit from enhanced controls and security protection.

It’s great to see Open AI providing dedicated models and instances for central and federal governments.

Microsoft confirm GPT-4o is now available on Azure AI

Just ahead of Microsoft Build, the Azure team have announced the availability of GPT-4o, OpenAI’s latest flagship model on Azure AI. This innovative multimodal model combines text, vision, and audio capabilities, establishing a new benchmark for generative and conversational AI experiences. GPT-4o is now available in the Azure OpenAI Service for preview, with support for text and image inputs.

This is a preview for testing now

What does GPT-4o Bring?

GPT-4o represents a paradigm shift in the interaction of AI models with multimodal inputs. It integrates text, images, and audio to deliver a more immersive and engaging user experience.

What does the “preview” include?

Currently in preview, Azure OpenAI Service customers will be able to test GPT-4o’s broad capabilities via a preview playground in Azure OpenAI Studio. This initial version emphasizes text and visual inputs, offering a preview of the model’s possibilities and setting the stage for additional functionalities, including audio and video.

The preview is free to try but has limitations around usage and location availability.

Designed for rapidity and efficiency, GPT-4o’s sophisticated processing of complex inquiries with fewer resources has the potential to offer both cost efficiency and enhanced performance.

Note: At time of writing, this is preview is available in two US regions only West US3 and East US.

What about GPT-4o in Microsoft Copilot?

We don’t know yet, but we do know that there will exciting updates around the rest of the Microsoft AI stack this week. Microsoft has an agressive and innovation fuelled roadmap for Microsoft 365 Copilot so as Microsoft continues to update and integrate OpenAI’s latest models into Copilot – I’m looking forward to hearing more this week.

What else is coming?

This week is Microsoft Build 2024 in Seattle and online. I expect this to be (pretty much) all about Copilot, and AI so expect to hear more about GPT-4o and other Azure AI updates.


Further Reading

You can read more about GOT-4o at the official OpenAI Blog which is < here >.

Prompt-a-long with Copilot in Word

Goal: Perfecting Prompting in Word

The goal of this blog post is to provide some field experience tips and coaching to help you get the best out of using Copilot in Word. For this you need to have either a Copilot Pro license or a Copilot for Microsoft 365 license and be signed into Word (or Word Online).

In this example, our Goal is to take a Marketing Analysis document we have been sent, and to draw out key information we can use in a “sales meeting” that we have coming up. The document contains lots of information relevant to different parts of the business.

I have provided a link to the document I used (courtesty of Microsoft) so you can either follow along or reference the videos included in this post. Welcome your comments – so please let me know how you get on.

Scenario: Using Copilot to pull key info from a Marketing Report

Instructions:

To work on this example with me, speak to your marketing team and obtain a Market Analysis Report for other similar document. For this example, I am using a public sample document Microsoft have shared called “Mystic Spice Premium Chai Tea.docx” which you can access here. The password for the link is “Copilot”. Once you have done that, save the file to your OneDrive so you can use it to test out and experiment with these prompts.

  1. Open the document you obtained (or use my link above) in Word and then
  2. Open the Copilot pane by selecting the Copilot icon in the top right of the “ribbon”
  3. Enter the prompts below and follow along.

The Starting Prompt

Enter the Starting prompt Summarise this Word Document” or click the suggested prompt to do the same thing…

In this simple prompt, we have started with what I call the “Alexa Prompt” – we are asking a simple question with a basic goal “to summarise the Word document”.

Using a standard/simple Copilot prompt to Summarise a Doc in Word.

This has done we asked but since we gave no context or information about what we wanted and why we needed it, Copilot has just read throught the document a pilled out key bits of information from each section.


The “better” prompt

To improve on this, we are going to repeat the prompt, but this time, we will add some more context to help Copilot understand the purpose of the summary and tailor the response for us accordingly.

Write a new prompt: “Summarise this document and create a brief overview of the main points to discuss with my team during the tomorrow’s Sales meeting“.

Here we are giving Copilot some more context specific about what information we want. It now knows why we need the information (for a sales meeting) and it knows to keep the ouput brief.

Using a more specific Copilot prompt to Summarise a Doc in Word.

If you run the prompt yourself (or check the video above) you will see that this time Copilot has pulled out specific around Market Trends and Demand inluding stats around CAGR. It has told us about the key competitiors, distribution and sales channels and also Sales Strategy, Outcomes and the Challenges in selling.


The Super Prompt

For the final prompt (I call it the “super prompt”), we are going to be even more specific with the ask to get just the information we need.

Use this prompt “Summarise this word document but focus on the Competitive Analysis section only. Provide a brief overview of the main points to discuss with my team during the tomorrow’s Sales meeting. Please keep the summary to 5 key points and use simple language.”

If you compare the output of this prompt to the previous ones, you will see that since we have been specific about where we want Copilot to focus, the response we get is specific to what we have asked. It’s still a summary, but it is focussed on the just the Competitor Analysis section and we have kept the response concise and in simple language. It knows to keep this simple and make it relevant to sales…

Follow along – or check out the video below where I run the prompt.

Our “Crafted” Super-Prompt in Copilot in Word

So there you go – we have started with a simple prompt and I have hopefully shown you that by thinking about what and why we want the information and also the audience the response is indended for, Copilot can produce information just the way we need it.

Summary and Lesson

So, I am pretty happy with that result. To recap – here is what we did to perfect our prompt…

  1. We started with our Goal (which was to summarise the document)
  2. Added some Context (that we want the information for a sales meeting)
  3. Specified the Source of the information (we asked to focus on competitive analysis), and
  4. Set clear Expectations, (we asked for five key points using simple language).
Our Final Prompt: 

"Summarise this word document but focus on the Competitive Analysis section only. Provide a brief overview of the main points to discuss with my team during the tomorrow's Sales meeting. Please keep the summary to 5 key points and use simple language".

This prompt has all the details it needs to give us the results we need. It has a Goal, Context, Source, and Expectations.

ChatGPT vs Bing Chat | Why you need to try Bing Chat!

Image of Bing Logo and ChatGPT logo

Everyone has been talking ChatGPT, but here I want to talk about the similarities, differences, and in my view, the HUGE ADVANTAGES OF BING CHAT OVER CHATGPT. Whilst both look similar in the way they work, are both powered by the OpenAI’s Generative AI technology and are conversational based, the speed, accuracy, quality of their responses and capability differs hugely.

Both Bing Chat and ChatGPT are based on the OpenAI GPT technology.

Natural language chatbot ChatGPT, quickly become the focal point in the tech industry since it was annouced and released to the masses at the backend of 2022. Whilst most initially saw it as a “fun way” to write poems, essays, answer exam questions and even generate computer code, 6 months later we are starting to see the true potential of AI as digitial assistants

We now have a plethora of AI tools, with new ones (it feels) coming every day. Outside of this fun stuff, the AI technology behind these tools is capable of so much more, and we are just at the beginning of what will change the way we interface with our devices and apps such as Microsoft Copilot – more on that here:

ChatGPT debuted in Nov 2022, and in Feb 2023, Microsoft their new AI powered version of Bing with Bing Chat being its’ standout feature, powered by the same technology (be it a Peter, better version) behind ChatGPT – a generative artificial intelligence algorithm created by Open AI. Recently Google had also released their AI chat tool Bard, which is not based on OpenAI GPT technology.

ChatGPT debuted in Nov 2022.

In Feb 2023, Microsoft their new AI powered Bing Chat, powered by the technology behind ChatGPT, a generative artificial intelligence algorithm created by Open AI.

Bing Chat vs ChatGPT – What’s the difference?

Bing Chat and ChatGPT both fall under the category of generative AI, which means they can generate brand-new text that has never been written before.

This kind of AI is useful since rather than leveraging keywords to look for text and return results based on a ranking, it works by providing the answer to the question you ask. Of course, it is only able to do this if it can provide accurate information. To do this is needs access to an entire world of data – like the Internet! So how does it work?

Both Bing Chat and Chat GPT are both built on a large language model (LLM), developed by OpenAI, of which Microsoft now own a majority share.

  • ChatGPT’s model, known as GPT-3.5, has been trained on billions of articles from across the internet but only as recent as end of 2021.
  • Bing Chat uses a newer, advanced, GPT-4 language model which can work with more different media types and not just text. Free to use.
  • ChatGPT offers a paid/premium version of its LLM (also using GPT-4), but it is only available today as a $20 per month subscription service.

The first major advantage of Bing Chat is that it, (since it’s also integrated into their native Bing Search, their Edge browser, and Windows 11), is that it can also search the internet for information and references in real-time. This is a huge advantage, because it’s up to date, but can lead also make it less accurate, since any misinformation it finds upon this search can introduce bias or error into its response. In most cases though, Bing Chat does reference its sources at the bottom of each response, which allows you to quickly verify the Bing’s accuracy and appropriateness by clicking on one of the citations.

Who “knows” more? Bing Chat vs Chat GPT

If you’re looking for information on a technical subject or recent event, Bing Chat will always deliver a higher-quality response as it can search the internet in real time so is “up to date”. ChatGPT does not do as well in recent events as it only has knowledge of events prior up to 2021.

This is the second big win for Bing, since it can be used as a copilot for the Internet, for anything from breaking news headline, shopping, days out, and travel since it does not have date limitations and can use the Internet to keep its LLM up to date.

The third key differentiator between the two is that Bing Chat lets you choose from three modes of conversation – Creative, Balanced, and Precise. Creative mode is closest to ChatGPT’s default behaviour.

  • More Creative: Longer, more descriptive, and “imaginative” answers.
  • More Precise: Shorter, straightforward search-focused answers.
  • More Balanced: Default mode – informative, yet friendly – middle of the other styles.

Visually Speaking | Bing Chat vs ChatGPT

The fourth standout feature for me is that you can use Bing Image Creator to generate images based on text prompts as pictured above. OpenAI does offer DALL-E for the same purpose. However, it’s not integrated within ChatGPT currently. What makes Bing Chat great is that you can use it in the flow of the conversation – or can use Bing Chat to create you an image on the fly. Bing uses also uses OpenAI’s DALL-E AI image generator (it is powered by Bing Image Creator) and has fewer limitations that the official OpenAI DALL-E website.

As shown below, by using the “More Creative” conversation style, you can start with a simple question and then elevate to an image all in the same context and without switching to a different tool.

Image showing text and image in Bing Chat

Where can you use it? Bing Chat vs ChatGPT

ChatGPT launched in late 2022 as a web-based tool and whilst there are a few third-party browser extensions the browser remains the main mode of use.

Bing Chat is free and available to all and accessible via the new Bing home page at https://www.bing.com.

The fifth standout feature for me, is all the other places you can access Bing Chat from. For example, there is a dedicated Bing app, you can access it via the Skype app, directly from the Edge Brower, from the Windows 11 search and via the SwiftKey keyboard app.

Current limitations | Bing Chat vs ChatGPT

Bing Chat is free and available to all and accessible via the new Bing home page at https://www.bing.com. as well as via Skype, Edge, Windows 11 and more. Bing Chat uses he newer, more powerful GPT-4 model and supports image search and image creation.

ChatGPT is also free but available only via the web and is limited to Text and uses GPT-3.5, though you can pay for the premium version ($20 a month) where you get higher-quality responses, thanks to the GPT-4 model and should also get priority access to the service, faster response times, and early access to new features. as they become available.

Bing Chat limits conversation length (currently) to twenty conversations per topic before you need to start again, meaning long back-and-forth conversations may not always be possible. Microsoft says that most people find what they’re looking for within five replies or fewer.

ChatGPT doesn’t have such restrictions and imposes very few restrictions on usage (even the free version), allowing for almost unlimited length conversations.

Bing Chat can understand most major languages including French, German, and Japanese.

ChatGPT will also respond in other languages, but the underlying GPT-3.5 model was primarily trained on English samples and text so responses in other languages are not as accurate.

Bing Chat does not impose limits on use and there is no premium tier. Responses are funded by advertising, meaning you may see ads that show up embedded inside the chat responses. For example, if you ask for holiday recommendations, you may see suggestions from local travel agents.

ChatGPT uses a token system, and you are limited to a set number of tokens per day for the free version. The premium (paid) version removes these limitations. Additionally, ChatGPT does not currently use an advertising model so is cleaner not biased.

So which one? Bing Chat or ChatGPT?

So, ok, I may be biased, but that doesn’t mean I am wrong, but in my experience, Bing Chat offers a far more premium offering over ChatGPT. It is free and is (if you like that), integrated across more of the applications, services, and platforms that you most likely use. My reasons to use Bing Chat are:

  1. Bing Chat can be accessed from Windows 11, Skype, Edge, Bing app and SwiftKey
  2. Bing Chat can search the web – it is natively integrated into Bing Search (the new Bing).
  3. Bing Chat lets you tailor your chat experience – providing a richer experience than ChatGPT.
  4. Bing Chat is not just about text chat, it can use images and even create AI powered images
  5. Bing Chat is uses GPT-4 and is still 100% free. ChatGPT 3.5 is also free but not as good and you have to pay $20 a month to use the better GPT-4 version. So just use Bing!

Questions and Answers

Is Bing Chat the “same” as Chat GPT?

No. While both leverage Open AI’s Chat GPT, Bing AI uses a more advanced model codenamed Prometheus that has more capabilities than ChatGPT. Prometheus is a proprietary technology from Microsoft that uses Bing and GPT to generate responses based on real-time data. Microsoft’s new Bing also leverages real-time information and OpenAI’s next-generation GPT-4 model to generate responses and can also search the internet. You cannot do this with ChatGPT, though there are browser extensions which provide some search functionality.

What does GPT stand for?

GPT stands for Generative Pre-Trained Transformer. “A GPT is a language model that has been trained on a vast dataset of text to generate human-like text. For example, the “Chat” part of “ChatGPT” refers to it being a chatbot”. [source: HowToGeek]

Who Owns ChatGPT?

ChatGPT was created by OpenAI, a San Francisco-based start-up. The same Large Language Model (LLM) also powers Microsoft’s Bing Chat. Microsoft owns a majority share in OpenAI and is also using the technology along with its propriety Prometheus model to power upcoming features in Microsoft 365 co-pilot.

What’s the difference between Bing Chat and ChatGPT?

Chat GPT uses GPT-3.5. Bing Chat uses, GPT-4 – OpenAI’s latest language model that’s smarter, safer, and more accurate. GPT-4 supports image inputs for the first time, allowing you to submit visual prompts like a drawing, graph, or infographic. GPT-4 is also smarter and more “creative”.

Chat GPT does have a version based on GPT-4 but is a paid-for premium service. Bing Chat uses GPT-4 already. Bing Chat is also integrated into Bing, Edge, Skype, and Windows 11.

What is Generative AI?

Put simply, Generative AI is a term used to describe computer algorithms that can generate text, images, videos, and audio all on their own based on natural language queries.

Before Generative AI, most AI systems weren’t highly creative and would deliver far worse results than a human. However, that’s no longer the case with generative AI. With Generative AI, you can ask an AI tool like Bing Image Creator to create a photorealistic image of a “cute blue AI creature with orange eyes” and it will deliver the results you see above. In this case, the AI has not been explicitly taught or trained to produce this image, it creates a unique image based on what it knows and what it has therefore “learnt” based.

As such generative AI is designed to mimic the way humans perform tasks. The first step is to extract patterns from existing data (the LLM), so if you want an AI that can generate a face for example, you will need a large dataset containing different images of faces. With enough training, the learning algorithm will learn what a face looks like as well as understanding the common features that a face has, such a nose, eyes ears, and lips. From here, it can start working on smaller details like expressions, facial hair, and skin tones to create unique images.

Image of people, created by Bing Images.

Summary

Hope you found this blog useful and informative and above it, that it’s made you want to try to Bing Chat. As a reminder, here’s how Bing Chat differs from ChatGPT.

  1. ChatGPT and Bing Chat both use GPT as their large language model (LLM), but Microsoft has adopted a more advanced model for Bing Chat which is more accurate and faster (currently) than Chat GPT.
  2. Since Bing Chat is built into Bing Search, the new Chat feature is more up to date that ChatGPT since it can also leverage web results to feeds it LLM. ChatGPT (currently) does not have knowledge of most recent events like up-to-date news.
  3. Bing Chat is available on more platforms than ChatGPT which seamless integration into Edge, Bing Search, Skype and even SwiftKey keyboard for iOS and Android.
  4. Whilst ChatGPT is great with text-based stuff, Bing Chat also has in-built AI image generators powered by Bing Image Creator (try it – it’s awesome).
  5. Bing Chat uses GPT-4 and is free whereas to get equivalent premium access to ChatGPT, you need to pay circa $20 a month which gives you priority usage and other benefits which include GPT-4 LLMs.
  6. Bing provides three different Chat modes (Creative, Balanced and Precise), while ChatGPT does not have any settings or fine-tuning its output and response mode.
  7. Bing Chat always provides links to its’ data sources whereas ChatGPT does not.
  8. Bing Chat supports many non-English languages and is trained on more than English. Whilst ChatGPT can understand non-English languages, its training was primarily on US English words and samples.