Amazon Invests $4bn In Anthropic, As ChatGPT Gets Voice, Image Features

Amazon said on Monday it would invest up to $4 billion (£3.3bn) into artificial intelligence firm Anthropic and take a minority stake in the firm, as rival OpenAI announced new voice and image features for its wildly popular ChatGPT.

Anthropic was founded by former OpenAI executives and makes a chatbot called Claude that competes with ChatGPT.

The company said under the new deal it would use Amazon Web Services as its primary cloud provider, and would provide AWS customers with early access to features for AI model customisation and optimisation.

Anthropic also said it would use AWS-designed semiconductors to train the foundation models that power its AI applications, in a move to demonstrate that Amazon’s custom AI silicon can provide an alternative to chips from market leader Nvidia.

Image credit: Levart Photographer/Unsplash

Amazon Bedrock integration

AWS will be given the ability to build on Anthropic’s AI models with Amazon Bedrock, a service that allows users to use existing models to build cloud-based generative AI applications.

The deal is one of many in which backers are pouring billions into the AI industry, following a multi-billion dollar investment by Microsoft in OpenAI in January.

Meanwhile OpenAI said in a Monday blog post that ChatGPT’s upcoming voice capability would “opens doors to many creative and accessibility-focused applications”.

The feature is to allow ChatGPT to narrate bedtime stories, provide information to users at the dinner table or speak text input from users.

ChatGPT voice and image features

The company said Spotify is using the new technology to provide translations of popular podcasts in the narrator’s own voice.

The image feature is to allow users to take pictures to “troubleshoot why your grill won’t start, explore the contents of your fridge to plan a meal, or analyse a complex graph for work-related data”, OpenAI said.

The feature competes with Google Lens, which also provides information from images uploaded by users.

The features are to be released to Plus and Enterprise subscribers in the next two weeks.

Matthew Broersma

Matt Broersma is a long standing tech freelance, who has worked for Ziff-Davis, ZDnet and other leading publications

Recent Posts

Spyware Maker NSO Group Found Liable In US Court

Landmark ruling finds NSO Group liable on hacking charges in US federal court, after Pegasus…

21 hours ago

Microsoft Diversifying 365 Copilot Away From OpenAI

Microsoft reportedly adding internal and third-party AI models to enterprise 365 Copilot offering as it…

22 hours ago

Albania Bans TikTok For One Year After Stabbing

Albania to ban access to TikTok for one year after schoolboy stabbed to death, as…

22 hours ago

Foldable Shipments Slow In China Amidst Global Growth Pains

Shipments of foldable smartphones show dramatic slowdown in world's biggest smartphone market amidst broader growth…

23 hours ago

Google Proposes Remedies After Antitrust Defeat

Google proposes modest remedies to restore search competition, while decrying government overreach and planning appeal

23 hours ago

Sega Considers Starting Own Game Subscription Service

Sega 'evaluating' starting its own game subscription service, as on-demand business model makes headway in…

24 hours ago