What Did OpenAI Do This Week? - 31/07/2023 [+25 LINKS]
CAN OPENAI BE FORCED TO OPEN UP?
A 157-page class-action lawsuit for $3 billion was filed in the Northern District of California against OpenAI this week. Launched by Clarkson, a California-based law firm that calls it out as “our chance to come together, to protect our identities and ideas, to put new safeguards in place, and to prepare for a future with AI”.
Clarkson has previously brought large-scale class-action lawsuits on issues ranging from data breaches to false advertising. Ryan Clarkson (Managing Partner) wants to represent “real people whose information was stolen and commercially misappropriated to create this very powerful technology”.
Clarkson’s hope is to have a court to institute guardrails on how AI algorithms are trained and how people are compensated when their data is used. The lawsuit uses reasonably strong language, claiming that such applications of AI could risk “civilizational collapse”, as OpenAI and other experts have warned. The suit also alleges OpenAI doesn’t do enough to make sure children under 13 aren’t using its tools; something that other tech companies including Facebook and YouTube have been accused of over the years.
Clarkson has anonymous plaintiffs, but he’s canvasing for more. In which case you’ve got to question how strong his floodgates are, given he could be referring to anyone who has ever published anything on the net, that isn't hiding behind a paywall. The $3 billion figure proposed in the damages statement then seems surprisingly low, even if it is a placeholder figure as actual damages will be determined if they succeed in court. The amount is based on the number of individuals potentially harmed by the harvesting of data, which it is suggested will be in the millions.
Legal sources aren’t clear on his chances of success. The issue surrounds the data pulled from the internet constituting ‘fair use’. Katherine Gardner, an intellectual-property lawyer at Gunderson Dettmer stated: ‘Artists and other creative professionals who can show their copyrighted work was used to train the AI models could have an argument against the companies using it, but it’s less likely that people who simply posted or commented on a website would be able to win damages’.
Oh, and Forbes loves them some WDODTW:
SO WHAT?
While this is likely to be a cash grab, the ‘gold rush’ is unlikely to be fast until OpenAI opens up (or is forced to open up) its black box to show where their data came from. Essentially, we don’t know what ended up in its LLM. What is rising is an urgency being placed on clawing back ownership of and citing the origins of digital content – whether labelling it AI-generated or protecting it from profiteering (see Reddit’s war and now Twitter is playing with access limits). What matters here is precedent; the lawsuit could force OpenAI to show the world how it works, but in reality, the legal battle to make this happen would likely be long, expensive and in both OpenAI and Microsoft’s best interests to delay and obfuscate as much as possible. Equally, showing how OpenAI works is near impossible given even the company still claims to not know how it works.
WANT TO INNOVATE LIKE OPENAI? Order your copy of the second volume of ‘Disruptive Technologies’ now!
OpenAI announced its first international expansion with a new office in London, United Kingdom. /OpenAI
OpenAI announced open roles for a Policy & Partnerships lead, Acct Executive (Go to Market) and Security Engineer (Detection & Response) in the new UK Office. /OpenAI
OpenAI received a letter from U.S. Senator Michael Bennet (on Thursday) urging them to label AI-generated content and limit the spread of material aimed at misleading users. /Reuters
Sam Altman privately told developers OpenAI wants to turn ChatGPT into a “supersmart personal assistant for work.” /The Information
ChatGPT guardrails and limits were set for Congress staff this week (no other language models authorized use). /Axios
OpenAI shuts Down its AI detection tool. /Decrypt
OpenAI released ChatGPT for Android in the US, Bangladesh, and Brazil, and plans to launch in more countries next week. /Verge
OpenAI was joined by Microsoft to help Moody’s develop a research tool to assess risk. /Bloomberg
ChatGPT partially powered the launch of booking.com’s trip planner in the beta test - US market only. /Reuters
ChatGPT iOS app added “Browse with Bing” as part of GPT-4. /Tech Crunch
OpenAI sued in San Francisco Federal Court by two U.S. authors (on Wednesday) claiming violated copyrights with misuse of their works (and others) to "train" ChatGPT. /Reuters
The Worldcoin Foundation rolled out its Worldcoin token, received after using an eye-scanning “orb”, in 35 cities; Worldcoin plans to retain 20% of its tokens. /FT
Sam Altman’s Worldcoin integrated with Okta. /CoinDesk
OpenAI is being sued in a 157-page class-action lawsuit for $3 billion, filed in the Northern District of California for allegedly stealing data from the public to train ChatGPT. /WP
Mira Murati earned a WFT Innovators Award. / Everydays.WTF
OpenAI paid engineers building ChatGPT “crazy money”. /San Francisco Standard
Junk websites filled with AI-generated text pulled in money from programmatic ads. /MIT Tech Review
An open letter proposing EU Artificial Intelligence legislation will jeopardise Europe's competitiveness and technological sovereignty was signed by 160+ executives at companies inc. Meta. /Reuters
OpenAI’s top rival ‘Inflection AI’ (personal AI) raised $1.3 billion from investors including Microsoft and Nvidia. /Reuters
Biotech begins human trials of a drug designed by artificial intelligence. /FT
Unilever's unveiled its ‘Generative AI Marketing Collective’ - an internal consortium helping it figure out how to apply AI creatively. /ADWEEK
Reid Hoffman spoke with Tyler Cowan about the Possibilities of AI. /Conversations with Tyler
Melinda French Gates is “very nervous” about how the current AI arms race and rush to implement AI into products will ultimately play out for women. /CNN
How AI’s deflationary winds will blow away profits. /Reuters