Since the emergence of generative artificial intelligence gained the spotlight following the launch of ChatGPT in 2022, the market has seen significant players enter the fray. Among these is Anthropic, which has made a name for itself among AI chatbots with its Claude AI.
In this article, we’ll examine how Anthropic has grown to rival large conglomerates like OpenAI and Google and how its AI differs from other models taking the industry by storm.
About Anthropic
Anthropic was founded as a startup in 2021 by former employees of OpenAI, the company behind the hugely successful ChatGPT. Within a year, the company secured over $500 million in funding from the crypto exchange FTX Trading, which later went bankrupt.
By the end of 2022, the first version of Claude was completed but was not released due to the company not feeling comfortable with the internal safety test results the AI had undergone.
A year later, after the first instances of Claude were released, Amazon gave Anthropic a major boost by investing $1.25 billion into the company, with a further $2.75 billion in investment paid in 2024. As part of the investment, the company agreed that all Claude services would be run using Amazon Web Services (AWS).
Despite competing with the company due to its own model (Gemini), Google also invested $500 million in Anthropic, with an agreement to invest an additional $1.5 billion later on. These injections of cash helped Anthropic grow substantially and produce outstanding AI models that offer a unique twist compared to those offered by Google and OpenAI.
AI Focus
One of the largest differences between Anthropic and competitor developers is the focus behind the AI models each has created.
Google’s focus on AI has been implemented through a range of applications and models, varying between the use of LaMDA (language model for dialogue applications) and PaLM (pathways language model). These two work hand in hand to create AI capable of open-ended conversations and models that can perform various tasks efficiently and achieve Google’s vision of creating sustainable AI with multiple uses.
OpenAI
Like Google, OpenAI is focused on creating large-scale models that offer human-like text responses, but it can also be applied to various industries for numerous purposes. In addition, OpenAI employs multimodal models like DALL-E to offer understanding not just of text but also the comprehension and generation of images.
Anthropic
In contrast to Google and OpenAI, Anthropic has an evident focus: ethical AI. This is achieved through the concept of constitutional AI, which forms the basis for all models created by the company.
Using a constitutional AI (CAI) framework, Anthropic has created Claude as the ultimate ethical AI. It aims to provide harmless and honest responses and assistance that don’t cross ethical lines.
To do this, the CAI uses a set of rules developed from sources such as the Universal Declaration of Human Rights to outline the desired behavior of how the AI model should work. In addition, it checks each response generated against these rules and then learns which responses are allowed and which should be reworked to better fit the ethical focus of the company.
Learning Methods
To ensure that Claude can learn to follow the CAI framework seamlessly, each version of the LLM (large language model) undergoes two levels of training. Both instances use AI to train themselves differently.
Supervised Learning
The first phase requires Claude to generate responses to specific prompts. These responses are then critiqued against the constitution provided by Anthropic. If the responses require revision, adjustments are made, and the model is fine-tuned to provide more accurate answers from the start.
Reinforcement Learning
After supervised learning is complete, the AI is required to generate responses to select prompts that are then compared against the constitution for compliance. Should any responses not be compliant, this information is then fed back into the LLM so that it can learn how to rectify instances of non-compliance in the future.
Anthropic’s Popularity
Aside from a focus on ethical AI, numerous factors have helped Anthropic grow in popularity to become one of the leading forces in the AI industry.
Approach to Safety
Anthropic’s focus and approach to the safe use of AI have attracted widespread attention and encouraged the use of AI by those who may not be guarded as significantly by other AI models.
An example is the AI’s refusal to provide information on specific topics, like drinking or online gambling. If somebody attempts to get the AI to suggest websites that are a great place for getting started when gambling online, it warns about the possible adverse effects of gambling rather than simply listing sources where the activity can be undertaken.
Natural Language
While all LLMs attempt to sound human when conversing, Claude excels at this and has been widely praised for its natural and contextual language. Because of how it is coded, Claude sounds more human than some competing AI models and can better understand context in conversation while offering more nuanced responses.
Versatility
Another reason Claude has grown to be so popular is its versatility. From creative and research writing to tutoring and coding, the LLM has a vast knowledge base that can be employed in many ways. This has made it a preferred choice among those with diverse AI needs.
Lack of Limitations
The final reason Anthropic has gained such a large user base is that the AI has fewer limitations than other options. Some of these limitations include that Claude does not require a sign-in, offers access to an advanced version of its LLM for free, and is able to retain a large conversation context for a more natural conversation flow.
