The battles for AI domination intensify
This week in AI: Google and Microsoft race to launch AI-powered search, Shutterstock and Getty Images respond differently to generative imagery, and real lawyers threaten legal action over the use of DoNotPay’s robo lawyer in a real-world courtroom.
While it seemed like everyone this week was on social media sharing news that ChatGPT passed both the bar and medical licensing exams, the big story was the fire lit under both Microsoft and Google to be first to the market with AI search.
Skip below for other AI news of the week. But first…
A WEST COAST RED ALERT
Google is reportedly on Code Red amid the exploding interest in ChatGPT, prompting a return of company founders Larry Page and Serge Brin. Last year, parent company Alphabet said it had 20 different AI tools in development. Now the pressure is on to actually get them to market, including its own search engine with a chatbot feature, this year.
If you think ChatGPT is impressive, you might have missed the news last summer that one of the Google engineers working on its AI described it as “sentient” and leaked a transcript of its conversation with researchers that blew my mind. (Google said the transcript was rearranged from nine different conversations for readability, but still, the responses to deep questions about self-awareness, emotions and fears, and concept of self was astounding.)
Meanwhile, Microsoft and OpenAPI announced its plans to continue its multi-billion dollar partnership. Then, Microsoft had its quarterly earnings call where CEO Satya Nadella told investors, “The age of AI is upon us, and Microsoft is powering it.” He added that Microsoft had the most powerful supercomputing infrastructure in the world. If it infuses ChatGPT functionality into its long-struggling Bing search engine, we could be looking at the tech comeback of the century.
DEVELOPMENTS IN TEXT-TO-IMAGE
It’s hard to imagine a future where any business or marketer would spend money on stock photography if they have the prompt-writing skills to create exactly what they want in Dall-e or Midjourney. Shutterstock seems to have split the difference, this week introducing its generative AI toolkit. But the catch is that the images are “ready for licensing” once generated. In other words, they still want you to pay.
They may be embracing AI more than Getty Images, which last week sued Stability AI over its Stable Diffusion AI art generation tool for using Getty images to train the model without permission. But Shutterstock is going to have to do better, as novice users can easily interact with the simple interface of Dall-e for free and more skilled and technical creators looking to generate higher quality digital art are flooding over to Midjourney.
One thing is certain, there will be more legal challenges ahead for ownership of generative content. Legal experts are divided about whether the practice of scraping the web for content to train AI models is covered by laws like the US fair-use doctrine or if this constitutes copyright violation. In another lawsuit to watch, a trio of artists last week filed a lawsuit against Midjourney, Stability AI and artist portfolio platform DevianArt, which recently created its own AI art generator, DreamUp.
For now, in a concession to artists, Stability AI says artists will be able to opt-out of the next version of Stable Diffusion.
A PIVOT AHEAD FOR ROBO LAWYER
The stage was set for the first AI-powered “robot lawyer” to actually defend someone in court next month, that is, until human lawyers got to lawyering. Reportedly, multiple threats of legal action and even prison time was too much for the British man who just wanted to have AI-powered DoNotPay help him beat a traffic ticket in California.
Read the article for how it was set to go down, why this approach is illegal in most courtrooms, why (and predictably) human lawyers are displeased, and how the company is pivoting amid threats from multiple state bar associations.
Most interesting to me as a marketing communications nerd is that CEO Joshua Browder is now reflecting on how positioning DoNotPay as a "robot lawyer" wasn't maybe such a good approach after all, since that language seems to have particularly riled up lawyers.
DoNotPay was originally built to contest parking tickets when it launched in 2015, but has really grown legs in the last several years. If you go to their website, you’ll see the insane list of legal issues it can tackle, from automatically cancelling your free trials to actually suing someone. The company says it has helped customers win more than 2 million customer service disputes.
It has raised $27.7 million from tech-focused venture capital firms, including Andreessen Horowitz and Crew Capital. So it seems that VCs are prepared to wade through the pushback and invest in the future of robot lawyers.
I'll keep my popcorn handy.
This article was written by Founder and Chief Elf Herder, Kimberly Morrison. As part of our commitment to transparency, we will aways disclose authorship of our blogs.
