These chatbots can’t keep their traps shut.
Anthropic, a California-based AI company, announced an agreement Monday with Amazon that will see the e-commerce giant invest up to $4 billion in Anthropic, starting out with a casual $1.25 billion. Amazon — along with every other Zuck, Musk, and Harry — is trying to get ahead in the generative AI race. It’s got its work cut out for it. On Monday, the company that was first out of the gate, OpenAI, said ChatGPT now has the ability to read aloud the text it spins up. Alexa, does that sound familiar to you?
By throwing its weight behind Anthropic, whose founders happen to include some OpenAI alumni, Amazon is deploying a similar strategy to Microsoft, which is a major backer of OpenAI. While all the Big Tech companies are keen to wet their beaks on the generative AI boom, Amazon’s rivalry with Microsoft is uniquely symmetrical in that both companies are major players in cloud services. Amazon is granting Anthropic access to some of its finer computing chips as part of the partnership, but the deal is not quite as iron-clad as Microsoft’s tie-up with OpenAI, which has an exclusivity deal for Microsoft servers. But AWS head Adam Selipsky seemed sanguine about that in an interview with the Financial Times, saying: “Exclusivity is not the goal.”
Amazon’s investment comes as the future of its smart assistant Alexa is under a slightly dark cloud. Last year, Amazon’s hardware division suffered as CEO Andy Jassy went on a cost-cutting mission. Rather than dream up new devices to sell to the Prime-subscribed masses, the company has signaled very clearly that melding generative AI into Alexa is its play:
- Dave Limp, Amazon’s outgoing head of its devices division, said at an event last week that the goal is to tone up Alexa’s chatting ability using large language models to make the assistant faster, more natural in its answers, and able to react to an interruption by its human masters. If anything’s going to precipitate the AI-human wars…
- Anthropic launched a ChatGPT-esque chatbot called Claude in March. Anthropic says Claude is a safer pair of hands than ChatGPT, claiming it’s less likely to come out with “harmful outputs.”
A Picture is Worth a Thousand Subpoenas: In the world of AI-generated images, a new line just got drawn in the digitally rendered sand. Getty, which in January sued AI company Stability AI on copyright grounds, announced it’s opening up its own AI-powered image generator and will absorb any legal fallout that might land on paying users who apparate images based on copyrighted material.