Welcome to the latest edition of FindBiometrics’ AI update. Here’s the latest big news on the shifting landscape of AI and identity technology:
British Prime Minister Rishi Sunak plans to propose the establishment of a global panel of experts to report on the state of artificial intelligence at a planned AI Summit next week. The panel would be modelled on the United Nations’ Intergovernmental Panel on Climate Change, which has reliably produced reports underestimating the impacts of climate change for decades now. High-profile leaders including European Commission head Ursula von der Leyen, US Vice President Kamala Harris, and US Secretary of Commerce Gina Raimondo are expected to attend.
The White House is expected to announce an Executive Order concerning artificial intelligence on Monday. The Washington Post reports that the order will require assessments of AI models before they can be used by federal agencies. It would also make it easier for highly skilled AI professionals to immigrate to the US.
Microsoft and Google heralded the rise of AI in quarterly updates. Microsoft CFO Amy Hood said that the company’s incoming business in its cloud operations was “primarily AI”, while Alphabet CEO Sundar Pichai called AI “a foundational platform shift” that would affect the tech giant’s entire business. Microsoft showed faster growth in cloud computing, and Wall Street responded, with roughly $100 billion in market value seeming to shift from Alphabet to Microsoft in the stock market.
Synthesia says almost half of the companies on the Fortune 100 list are its customers. The UK-based startup offers software enabling the creation of realist-looking AI avatars that can be made to speak any lines of text input. Users can choose from a cast of more than 140 ready-made avatars, or make deepfakes of themselves.
Stability AI’s general counsel and head of human resources have both departed the company. It may be an ominous sign for the Silicon Valley startup, which is best known for its Stable Diffusion image generation technology.
Shutterstock has integrated image generation tools into its popular stock image platform. The tools, which are still going through beta testing, allow users to edit existing images, for example by removing an object within a picture and typing in what they’d like to see in its place. The service is currently relying on DALL-E 2, the previous version of OpenAI’s image generation system.
Xiaomi appears to be building a deepfake generation tool directly into its forthcoming HyperOS operating system. Given enough photos of an individual to train on, the tool will allow the user to generate photos of that person in different settings. Presumably, the company envisions users faking a trip to Paris and the like, but it isn’t yet clear how the system would prevent them from generating images of other people, such as celebrities or politicians.
The chatbot’s take: We wanted a little bit of analysis about Stability AI’s executive changes:
October 27, 2023 – by Alex Perala