GPTBARD

The Future of Chatbots: Examining the Competition Between ChatGPT and Google Bard

July 07, 2023 / Katarina Rudela
Reading Time: 10 minutes

Software technology continues to develop rapidly in many areas, including artificial intelligence (AI). The increasing use of AI has resulted in the emergence of chatbots like OpenAI’s ChatGPT and Google’s Bard, which are designed to provide human-like responses. ChatGPT is a robust solution with many features and services that’s already finding many use cases. Furthermore, it’s currently undergoing significant changes, particularly with respect to its AI.

Google has recently launched Bard to compete in the chatbot market, especially against ChatGPT. Bard’s availability to the public will increase in the following weeks, which is likely to dramatically affect this sector. In particular, the capabilities of chatbots will grow quickly as these two solutions struggle for market share. A critical overview of the similarities and differences between ChatGPT and Bard is thus essential for understanding how they will improve user experiences.

Market Growth

The chatbot market is expected to grow from a market value of over $250 million in 2017 to more than $1.34 billion by 2024. A study by Global Market Insights provides additional details of the market growth in this sector during this period:

Chatbot Market

Figure 1: Chatbot Market Growth through 2024

Chatbot Market Growth through 2024

The market growth predicted above is based on the increased use of chatbots, especially for content marketing activities like advertising and digital marketing. The driving factors include technological developments that provide greater customization, which is one of the most desirable features of chatbots. The personalized capabilities of the latest chatbots better enable them to assist, engage and interact with customers.

Chatbots have been able to gain insights into customer behavior based on their responses since about 2016, thus improving their experience in making purchases. In addition, investors are providing the funding that startups need to develop chatbots with natural language processing (NLP) capability. These investments create marketing and promotional opportunities for the platforms that improve purchases and other transactions, along with customer service. At the same time, security issues and a lack of awareness regarding the benefits of chatbots among businesses is hindering growth.

Customer service applications currently account for the largest share of the chatbot market at 43 percent. The next largest shares in descending order are held by social media, order processing and marketing applications. Customer service is also the most important basis for competition in mature markets, where price and quality are similar among competitors. Intelligent chatbots allow organizations to improve their customer service by providing 24/7 support, including real-time queries and customer education.

E-commerce will generally derive the greatest benefit from the chatbot market, especially as the capabilities of AI develop over time. The banking, financial services and insurance (BFSI) sector also provides an opportunity for growth in chatbots, where they can help customers make better financial decisions, improve customer experience and prevent fraud.

The web-based chatbot market is expected to grow at an average Compound Annual Growth Rate (CAGR) of 36.9 percent between 2018 and 2024, which is a high rate to maintain for six years. This growth will be most applicable to companies selling products directly from their own websites. The AI-based chatbot market should grow at an average CAGR of 53 percent during the reporting period, as it provides customers with conversational experiences via voice and text. Increased collaboration among AI-based companies will also drive the growth in this market.

The majority of the chatbot market is held in North America, where tech companies are investing heavily in the research and development (R&D) needed to implement new technologies that will improve user experiences. Other organizations are also forming strategic partnerships and making acquisitions that will gain market share throughout the world. For example, the Asian Pacific region is expected to show the highest growth rate during the reporting period at 36 percent. This growth will be largely due to the mainstream use of chatbots as they become increasingly important in many industries. In addition, competition in the global chatbot market will get stronger as multinational firms and startups get more involved in R&D.

What is ChatGPT?

ChatGPT is a chatbot that OpenAI developed and released as a prototype in November 2022. ChatGPT quickly gained attention due to its articulate, detailed responses across many bodies of knowledge. Its biggest drawback is it’s inconsistent factual accuracy. OpenAI was valued at $29 billion following the ChatGPT’s release, effectively doubling its net worth. The following chart provides examples, capabilities and limitations of ChatGPT:

ChatGPT Examples, Capabilities, Limitations

ChatGPT vs Google

Note the limitation regarding world events that occurred after 2021, which is when ChatGPT’s formal training ended. Google will need to overcome this restriction in some way to remain competitive over the long term.

Training

ChatGPT is built onto OpenAI’s existing GPT-3 family of language models, although it has been further tuned by using reinforcement and supervised techniques. Both methods used human trainers to improve its performance, in collaboration with the Microsoft Azure supercomputing infrastructure.

The reinforcement stage used trainers to rank responses that ChatGPT made in previous conversations. These rankings help create reward models that further refined the responses based on Proximal Policy Optimization (PPO). The PPO algorithms provided a cost-effective alternative to trust-region policy optimization by eliminating the need for some computationally intensive algorithms. The supervised learning stage involved the trainers providing conversations as both users and AI assistants.

After the initial training period, OpenAI has continued to gather data from users, allowing it to further fine-tune ChatGPT. Users can vote for or against its responses, and also provide more detailed feedback. The goal of this process is to create sophisticated writing that appears to be written by a person.

Technological Impact

The large language model that ChatGPT currently uses is only the most recent in a series that OpenAI has produced. The ability to effectively interact with users in a variety of languages is one of this model’s latest features, which has attracted widespread interest while generating controversy over the ethics of this technology. ChatGPT will greatly affect the scientific community, especially for its ability to perform linguistic tasks, like NLP, translation and textual analysis.

Many users are already leveraging big language models to write various types of content, including essays, presentations and reviews. They’re also using it revise existing material, identify areas requiring further research, analyze statistics and write programming code. This technology will eventually be able to compose articles from scratch, plan experiments, conduct peer reviews and make editorial judgments. However, the biggest drawback to big language models will continue to be their tendency to generate illogical or biased content.

What is Bard?

Google created Bard in direct response to ChatGPT, particularly its potential competition with Google Search. Upon ChatGPT’s release, Google CEO Sundar Pichai quickly assigned multiple teams to further his company’s AI efforts. Google’s co-founders, Sergey Brin and Larry Page, also came out of semi-retirement to discuss the best response to ChatGPT.

In January 2023, Google began extensively testing competitors for ChatGPT, including one with the code name “Atlas.” It chose “Bard” as the official name for this product, due to the creative nature of its underlying algorithm. On February 6, 2023, Google announced the release of Bard, a conversational AI-based chatbot. It was initially released to a group of trusted testers, with a general release scheduled for the end of that month.

A livestream demonstrating Bard’s abilities on February 8 proved disappointing, and the YouTube video of it has since been made private. The next day, Google’s stock fell eight percent, equivalent to $100 billion in market value. Viewers were particularly critical of the incorrect information that Bard provided regarding the James Webb Space Telescope. Even Google’s own employees have described Bard’s release as “botched” and “rushed” in internal correspondence.

Analysts have pointed out that Bard’s release date was no coincidence, since Microsoft had planned to announce its partnership with OpenAI on February 7. The software giant has previously invested $1 billion in OpenAI, and this latest deal will provide OpenAI with $10 billion in R&D funding over the next few years. Microsoft’s motive for investing in ChatGPT is to integrate it into its Bing search engine. Another factor influencing the timing of this activity is that Microsoft and Google agreed to a truce in R&D on search engines in 2015, which expired in 2021.

Technology

Bard is based on Language Model for Dialogue Applications (LaMDA), Google’s international AI tool. LaMDA uses a transformer language model that uses a text corpus of 1.56 trillion words that describe dialogs and documents. Manually annotated responses then fine-tune Bard to improve its sensibleness, interestingness and safety. Bard attempts to integrate human knowledge into this language model to provide brief, inventive responses. Its explanations are designed to be understandable by anyone, even on complex topics like space discovery.

It leverages information throughout the world and should improve as users begin providing feedback on its responses. In the meantime, Google will be adding more routine duties to Bard’s capabilities. LaMDA has allowed this chatbot to become competitive with ChatGPT in a short period of time, despite its disappointing start. As a result, Bard should start to take market share from ChatGPT over the next few years.

Differences between Bard and ChatGPT

Bard and ChatGPT use very different approaches to achieve the same goal, providing human-like responses to questions on a variety of topics. The types of questions the user is asking, and the level of detail needed in the responses are crucial factors in determining the best chatbot. The following chart summarizes the key differences between Bard and ChatGPT:

Bard Vs ChatGPT

Figure 3: Bard vs ChatGPT

Bard vs ChatGPT

Bard uses the LaMDA language model, while ChatGPT uses GPT-3. Bard is closer to a traditional search engine than ChatGPT, which isn’t surprising given that Google also makes the world’s leading search engine. ChatGPT is more of a language-based technology that’s better designed for creating scripts, composing emails and completing other writing assignments. The ethics of using ChatGPT is currently a matter of hot debate, although no general consensus on this issue has been reached yet.

Bard uses the internet as its data source, which is an extremely large amount of existing data. ChatGPT gets its information from a customized data feed that OpenAI had to create. As a result, Bard has many sources of information that allow it to verify its content, while Chat GPT has only one source of information. This difference will provide Bard with an advantage over ChatGPT, once Bard is incorporated into Google’s search engine. Bard will thus have the edge over ChatGPT with regard to the scope and extent of the information it can access.

Bard’s information will remain current since it comes from the internet, which is always receiving updates. ChatGPT’s training ended in 2021, so its knowledge is restricted to events that occurred no later than that year. Bard’s information should get more reliable over time automatically, while ChatGPT will eventually require additional updates.

Bard’s access is limited at this time because it hasn’t been released to the general public yet, although this should occur in the near future. ChatGPT is already available to the public, so many people are already using its free version. Once Bard is incorporated into the Google search engine, it will be easy for anyone to access. ChatCPT requires users to create an account and log on before they can access it. Bard is currently free to use, as is ChatGPT at this time. However, ChatGPT will probably require a fee at some point.

Bard’s biases are based on the internet, while ChatGPT’s biases are based on its data feed. Bard provides users with more reliable information, and is also proficient at breaking up complex ideas into smaller pieces that are easier to understand. It can provide the user with many possible answers, allowing it to cover all aspects of the topic. This capability allows Bard to disseminate information more widely, especially to children. ChatGPT is more prone to making factual errors, but is better able to produce relevant answers in response to a prompt. On the other hand, ChatGPT’s answers tend to be more clear and unambiguous, as if they were coming from a friend.

Summary

The end of the “cease fire” on search engines between Google and Microsoft signals the beginning of an AI technology race between the two software giants. Smart devices like virtual personal assistants already use AI technology, which will change the way we live and work. However, it’s still unclear which of these models will prevail at this time. The use of AI is also controversial due to the possibilities of job loss, at least in the short term. In addition, AI chatbots also raise ethical concerns due to their potential for writing content without direct human intervention.

The release of Bard will significantly accelerate the shift towards AI technology, as it begins to compete with ChatGPT. Both chatbots will gain new features to help them achieve dominance in this space, especially with regard to AI capabilities. This competition will drive chatbots to large-scale operations, which they’ll need to receive widespread use by machines and people. In particular, Bard will need to improve its NLP, while ChatGPT will need a way to update its database with new events.

About Baytech

Baytech is passionate about the technology we use to build custom business applications, especially enterprise solutions that optimize business processes. We’ve been delivering software solutions in a variety of technologies since 1997. Our success is due to the skill and efficiency of our senior staff, which includes software engineers, project managers, and DevOps experts. All of our engineers are onshore, salaried staff members.

We focus on the quality, usability, and scalability of our software and don’t believe in mitigating cost at the risk of quality. We manage project costs by implementing an efficient development process that’s completely transparent and uses the latest standards and practices to build software right the first time. Contact us today to learn more about how we can help your business. Find us online at https://www.baytechconsulting.com/contact.