Dario Amodei, co-founder and CEO of AI startup Anthropic.
Chesnot | Getty Images
Anthropic, the Amazon-backed AI startup founded by former OpenAI executives, announced Tuesday that it has reached a milestone in the company's artificial intelligence space: artificial intelligence agents that can use a computer to complete complex tasks like a human.
Anthropic is the company behind Claude – a chatbot that has become very popular, like OpenAI's ChatGPT and Google's Gemini. Startups like Anthropic, along with tech giants like Google, AmazonBoth Microsoft and Meta are part of an AI arms race to ensure they are not left behind in a market expected to reach $1 trillion in revenue within a decade.
Anthropic's new computer capability, part of its two newest AI models, allows its technology to interpret what's on a computer screen, select buttons, enter text, navigate through websites, perform tasks through any software and browse the Internet in real time.
The tool “can use computers the same way we do,” Anthropic's chief science officer, Jared Kaplan, told CNBC in an interview, adding that it can perform tasks in “dozens or even hundreds of steps.”
Amazon had early access to the tool, and early customers and beta testers included Asana, Canva and Notion, Anthropic told CNBC. The company has been working on this tool since early this year, according to Kaplan.
Anthropic released the feature on Tuesday in public beta for developers. The team hopes to make it available to consumers and enterprise customers within the next few months, or early next year, according to Kaplan.
Future consumer applications include booking flights, scheduling appointments, filling out forms, conducting online research and submitting expense reports, Anthropic said.
“We want Claude to be able to actually help people with all different types of work, and we think the chatbot setup is somewhat limited because you can ask a question and (get) context but it stops there,” Kaplan told CNBC.
What is an AI agent?
Following the huge popularity of OpenAI's ChatGPT software, the industry quickly moved from previous text responses to AI-generated images, videos, and audio. Now, startups and big tech companies alike are turning to AI clients.
Rather than simply providing answers — the world of chatbots and image generators — agents are designed for productivity and to complete complex, multi-step tasks on behalf of the user. Although the term has not been precisely defined across the technology sector, AI agents are seen as a step beyond chatbots, as they are typically designed for specific business functions and can be customized on large AI models. Think JARVIS, the multi-faceted AI assistant to Tony Stark from the Marvel Universe.
There has been a “significant increase” in interest among technology investors in startups focused on building AI agents, Grace Esford, a partner at investment firm Lux Capital, told CNBC in June. They have collectively raised hundreds of millions of dollars and seen their valuations rise alongside the broader generative AI market.
Microsoft CEO Satya Nadella said on an earnings call earlier this year that he wants to introduce an AI agent that can complete more tasks on behalf of the user, though there is “a lot of implementation to come.” Executives from Meta and Google also praised their work in pushing AI agents to become increasingly productive.
Anthropic competes with OpenAI on multiple fronts
Anthropic has become one of the hottest AI startups since it released the first version of Claude in March 2023, a product that competes directly with OpenAI's ChatGPT in both enterprise and consumer markets, without any consumer access or much fanfare. Backers include Google, Salesforce, and Amazon, and since January, it has offered iOS and Android apps, a group plan for businesses, and international expansion into Europe.
“(We're) moving to a world where these models behave more like virtual collaborators than virtual assistants,” Scott White, product manager at Anthropic, told CNBC in September.
Anthropic's announcements Tuesday are the latest step in its long-term strategy to build those virtual collaborators, or agents.
Last month, Anthropic introduced Claude Enterprise, its biggest new product since its chatbot debut, designed for companies looking to integrate Anthropic's AI. Beta testers of the enterprise product and early customers include GitLab, Midjourney and Menlo Ventures, according to the company.
Claude Enterprise allows customers to upload relevant documents with a much larger context window than before — the equivalent of 100 30-minute sales conversations, 100,000 lines of code or 15 full financial reports, according to Anthropic. The plan also allows for “activity feeds” of power users within the company to show those who are newer to the AI field how others are using the technology, White said.
The launch of Claude Enterprise follows Anthropic's June debut of its more powerful Claude 3.5 Sonnet, and its May launch of its “Team” plan for small businesses.
In June, Anthropic also announced “Artifacts,” which it said allows a user to ask the Claude chatbot, for example, to create a text document or code and then open the result in a custom window.
White told CNBC in September that the artifacts, or “workspaces” that allow users to “see, edit and build on Claude’s creations in real time,” will allow Anthropic’s enterprise-level customers to create marketing calendars, feed sales data, create dashboards or… forecasts, draft feature code, write legal documents, summarize complex contracts, automate legal tasks and more.
Shortly after Anthropic debuted Teams in May, Mike Krieger, co-founder and former CTO of Meta-owned Instagram, joined the company as chief product officer. Under Krieger, the platform has grown to 1 billion users and its engineering team has grown to more than 450 people, according to a press release. Jan Lek, former safety lead at OpenAI, joined the company that same month.
