Amazon-backed Anthropic’s AI tool Claude aims to use computers like humans

Anthropic said the new version of its AI model Claude can now potentially emulate the way people interact with their computers.

Anthropic — which is backed by Amazon (NASDAQ:AMZN) and Alphabet’s (GOOGL) (NASDAQ:GOOG) Google — said the latest version of Claude 3.5 Sonnet can follow a user’s commands to move a cursor around their computer’s screen, click on relevant locations, and input information via a virtual keyboard. Thus emulating the way people interact with their own computer.

The skill is currently in public beta mode.

The company trained Claude to interpret what is happening on a screen and then use the software tools available to carry out tasks.

Anthropic noted that computer use is an entirely different approach to AI development. Until now, large language model, or LLM, developers made tools to fit the model, producing custom environments where AIs use designed tools to complete tasks. “Now, we can make the model fit the tools — Claude can fit into the computer environments we all use every day,” said the company in a press release.

The AI startup — which is on track to generate $1B in revenue this year — said its goal for Claude is to take pre-existing pieces of computer software and simply use them as a person would.

Anthropic found that the Claude 3.5 Sonnet, including its new computer use skill, remains at AI Safety Level 2 — which means that the AI model does not need higher standards of safety and security measures than those the company currently has.

One concern the company identified was ‘prompt injection’ — a type of cyberattack where malicious instructions are given to an AI model, causing it to either override its prior directions or perform unintended actions which deviate from the user’s original intent.

Anthropic said that since Claude can interpret screenshots from computers connected to the internet, it is possible that it may be exposed to content that includes prompt injection attacks.

Anthropic acknowledged that there is still a lot to do. “Even though it’s the current state of the art, Claude’s computer use remains slow and often error-prone,” the

Anthropic’s AI product Claude has grown in popularity. Amazon’s updated Alexa, reportedly, will be powered mainly by Anthropic’s AI models Claude rather than the tech giant’s own AI,

Claude, amid the crowded LLM market, competes with OpenAI’s ChatGPT for both enterprise and consumer markets. Meanwhile, other tech heavyweights, such as Microsoft, are also trying to woo enterprise customers with their AI offerings.

On Monday, Microsoft announced that it would allow businesses to start making their own autonomous AI agents from next month. Rival Salesforce (CRM) had already introduced its own configurable AI models in September.