The past week saw a flurry of activity in the world of Artificial Intelligence, with multinational conglomerates making significant decisions regarding their AI aspirations and contributions to next-generation technology.
Furthermore, there were notable developments in the political arena as top executives from major AI companies, including Microsoft, OpenAI, and Anthropic, held talks with Vice President Kamala Harris and President Biden at the White House.
Here’s a roundup of the latest AI news from the week.
White House Hosts Meeting with Top AI CEOs
The White House recently held a meeting to discuss the risks and safeguards of generative AI, with the participation of CEOs from major companies in the field such as Google, Microsoft, OpenAI, and Anthropic.
President Biden, who has personally experimented with ChatGPT, emphasized the importance of mitigating the current and potential risks that AI poses to individuals, society, and national security.
The meeting was primarily focused on companies that are currently leading in the space, particularly those with consumer-facing products. Notably absent from the meeting was Mark Zuckerberg, the CEO of Meta, as he was not invited by the White House.
Geoffrey Hinton, AI Pioneer, Quits Google and Warns of AI Risks
Geoffrey Hinton, a prominent figure in the field of artificial intelligence and a key player in the development of neural networks and deep learning, has announced his departure from Google. In a surprising move, he spoke out about the potential risks of AI and expressed regret for his work on the technology that paved the way for systems like ChatGPT.
Hinton cautioned that AI chatbots could surpass human intelligence and could potentially spread misinformation, manipulate public opinion, and pose a significant existential threat. He further expressed concerns about Google’s role as a responsible AI steward and pointed to competition from Microsoft’s Bard as a cause for concern.
Salesforce Unveils SlackGPT, Bringing AI Capabilities to Slack
Salesforce has recently introduced SlackGPT, an innovative generative AI experience designed for Slack. With SlackGPT, users can easily integrate and automate their preferred languages models like ChatGPT or Claude through partner-built apps or custom integrations.
This new tool also offers built-in AI features within Slack, such as conversation summaries and writing assistance, enhancing the communication experience for users. Additionally, SlackGPT introduces the new Einstein GPT app that leverages AI-powered customer insights from Salesforce data.
The primary objective of SlackGPT is to empower users to work more efficiently, learn faster, and communicate more effectively by integrating trusted generative AI capabilities into their daily work routine.
Midjourney has released version 5.1
Midjourney has unveiled version 5.1 of its generative AI art service, promising improved accuracy, fewer text artifacts, and increased sharpness. The latest release is more opinionated, bringing it closer to version 4’s results but with higher quality.
Midjourney 5.1 also introduces a “raw” mode for users who prefer images that are less opinionated. A test of the new version using identical prompts in both versions 5 and 5.1 highlights the significant enhancements in quality and dynamic compositions.
Midjourney customers can switch to version 5.1 using the /settings command in Discord, with plans to make it the default mode by the end of the week.
Box AI brings advanced AI models to Box Content Cloud
Box, the cloud content management solution, has announced the upcoming release of Box AI, which integrates advanced AI models from OpenAI with Box Content Cloud. This integration provides users with a suite of AI capabilities to help them quickly find relevant information within an organization’s files, making organizational files more valuable and useful.
Box AI will also allow users to generate content directly from Box Notes using existing data. The technology is governed by Box’s built-in permissions, ensuring it complies with enterprise-grade security, privacy, and compliance standards. The possibilities for improved workflows across various departments are significant, from financial analysis to HR and legal teams.
MIT develops AI-based tools to enhance pharmaceutical manufacturing and drug discovery
Researchers from MIT and Takeda have collaborated to develop two novel AI-based models that could significantly improve pharmaceutical manufacturing and drug discovery. The first model, called PEACE, utilizes AI to characterize rough particle surfaces in pharmaceutical pills and powders, enabling the optimization of the drying process and reducing the number of failed batches and experiments.
The second model, EquiBind, utilizes geometric deep learning to predict the best ligands for protein targets, enabling it to find potential drug molecules a thousand times faster than existing models.
These AI innovations have the potential to accelerate drug discovery and development, increasing efficiency, lowering costs, and improving success rates.
Bing’s ChatGPT-Enhanced Search Now Available to Everyone Without Waitlist
Microsoft has made its AI-powered search engine, powered by OpenAI’s ChatGPT, freely accessible to the public, removing the waitlist previously required. Along with this change, Microsoft has also made its AI-enabled Edge web browser available for free.
The new Bing search engine comes with features such as a chat history that allows the AI to remember previous conversations, visual elements like charts and graphs, and an expanded Bing Image Generator that works with over 100 languages.
While the move is undoubtedly a business strategy to onboard new users, it makes chatbots easily accessible and could transform the way search engines are used.