The Gist

  • AI testimony. OpenAI CEO, Sam Altman, discusses AI safety and ethics before Senate Judiciary Committee.
  • AI misinformation. AI expert Gary Marcus highlights potential risks of AI generating extreme misinformation.
  • AI investment. Amazon Web Services commits $12.7 billion to bolster India's cloud infrastructure by 2030.

On Tuesday, members of the Senate Judiciary Committee heard testimony from prominent figures in the AI industry, including OpenAI CEO Sam Altman, in a hearing focused on addressing issues in emerging AI technology.

Altman gave an overview of his company, its products and addressed OpenAI’s safety and ethical policies, testifying that, “While we believe the benefits of the tools we have deployed vastly outweigh the risks, ensuring their safety is vital to our work and we make significant efforts to ensure safety is built into our systems at all levels.”

He also offered an example of how AI is benefiting society by sharing the story of Ben Whittle, a pool installer and landscaper with dyslexia, who Altman said now benefits from an AI tool built on top of GPT-tech, to help him better communicate with clients.

Also appearing before the committee were Christina Montgomery, chief privacy & trust officer at IBM and scientist; and AI entrepreneur and New York Times best-selling author, Gary Marcus, who began his testimony with a bit of an alien-invasion eye-opener.

“In a shocking revelation, a cache of classified memos and documents leaked on the Discord ... has ignited a firestorm [suggesting] the U.S. Senate is secretly manipulated by extraterrestrial entities,” he said, before quickly explaining that the information came from GPT-4, and he shared it to highlight the potential for extreme misinformation.

“There are other risks, too, many stemming from the inherent unreliability of current systems. A law professor, for example, was accused by a chatbot that claimed falsely he committed sexual harassment — pointing to a Washington Post article that didn’t exist,” Marcus said. “The more that happens, the more anybody can deny anything. As one prominent lawyer told me Friday, defendants are starting to claim that plaintiffs are ‘making up’ legitimate evidence. These sorts of allegations undermine the ability of juries to decide what or who to believe … and contribute to the undermining of democracy.”

The hearing was part of a mounting and ongoing effort from the administration to draft legislation addressing AI technology.

In other AI news...  

Amazon Launches Bedrock, a New Fleet of AI Tech

Amazon recently unveiled its new fleet of AI tech called Amazon Bedrock. Amazon Bedrock is a service offered by Amazon Web Services (AWS) that provides users with pre-trained machine learning models (FMs or Functioning Models) from prominent AI startups as well as Amazon's own. These models can be accessed through an API (Application Programming Interface), allowing you to choose the best-suited model for your specific requirements.

As a "fully managed service," it appears that AWS handles all the underlying infrastructure, including servers, storage and networking, allowing users to focus on using the models rather than worrying about maintaining the infrastructure. And Bedrock's "serverless experience" means you don't have to deal with any server management tasks, so users can quickly start up the service, customize the pre-trained models their own data, and incorporate them into their applications. AWS tools and capabilities, including integration with Amazon SageMaker machine learning features, enable the ability to test different models (through SageMaker Experiments) and manage models on a large scale (through SageMaker Pipelines).

Related Article: Exploring the Game-Changing Potential of AI: Insights and Tips From Expert Daniel Wu

SAP Unveils New AI Capabilities, Including SAP Digital Assistant

This week during the SAP Sapphire event & ASUG Annual Conference in Orlando, SAP announced an expanded generative AI partnership with Microsoft and unveiled more than a dozen new business AI capabilities — including SAP Digital Assistant, available later this year, and aimed at improving customer experience with a unified natural language interface for SAP solutions.

“SAP Digital Assistant is also expanding its capabilities to provide users with proactive and contextual assistance inside their applications, starting with SAP Customer Experience solutions,” company officials said in a statement. “With the new SAP Digital Assistant for customer experience, sellers and support teams can easily get holistic insights, create recommendations, and automatically generate content to personalize customer experiences.”

Other new AI tech announced includes new generative AI capabilities in SAP SuccessFactors Recruiting that enable hiring managers to create more compelling, accurate job descriptions and SAP Transportation Management to enable automotive and manufacturing companies to automatically process more data in various formats.

Learning Opportunities

Related Article: Ex-Googler Godfather of AI: Machines, AI on Dangerous Path

Stability AI Introduces StableStudio

Stability AI, an artificial intelligence firm, just launched StableStudio, an open-source version of their acclaimed text-to-image consumer application, DreamStudio.

Following the launch of StableDiffusion in August, DreamStudio has been the main platform for introducing new models and features by Stability AI. The latest enhancement involved integrating SDXL, the company's most recent image generation model, into DreamStudio.

Initially conceived as an “animation studio for Disco Diffusion” the company’s focus shifted toward image generation with the arrival of Stable Diffusion in the summer of 2022.

“Our goal of creating a great multi-modal experience for generative AI has endured. To that end, we’ll be working on releasing our upcoming chat interface within the StableStudio project,” company officials said in a statement. “From enabling local-first development to experimenting with a new plugin system, we’ve tried hard to make things extensible for external developers.”

Other new features on the horizon include local inference through WebGPU and stable-diffusion-webui, desktop installation and ControlNet tools.

AI Video of the Week: Examine AI for the Sake of Our Collective Future

Just before testifying on Capitol Hill last Tuesday, AI researcher Gary Marcus called for an urgent reevaluation of whether we're building reliable AI systems — or misinformation machines, during a TED Talks event on May 12. Take a look.

AI Tweet of the Week: Artificial Intelligence Takes the Stage

In support of the current writer’s strike, most major late-night TV talk shows have stopped live shows in favor of reruns. Curious Refuge, an AI filmmaking company that offers videos, blogs and interactive courses for others interested in the craft of AI film, recently tweeted a video of a late-night talk show-inspired sketch. They say was created using only artificial intelligence, with all the news curation, jokes, artwork and voice done completely with AI.

Is it funny? Eh... But you be the judge.