In partnership with

Welcome to The Dispatch! We are the newsletter that keeps you informed about AI. Each Thursday, we aggregate the major developments in artificial intelligence; we pass along the news, useful resources, tools or services, and exciting projects in open source. Even if you aren’t an engineer, we’ll keep you in touch with what’s going on in AI.

And after a wild first year, things don’t appear to be slowing down any time soon. More rumors have surfaced in the last week about Q* (“Q-star”), a mysterious breakthrough project that may have been a driving factor in the company’s highly publicized recent drama. The project’s name implies it could be a combination of two well-known AI techniques, Q-learning and A* search - rules that can help the AI reason in a more human-like manner (understanding things “out of experience”, rather than just regurgitating training data) and lead to reduced hallucinations and increased mathematical capabilities.

Such a breakthrough would be a big step towards OpenAI’s goal of creating superintelligence. Newly reinstated CEO Sam Altman confirmed the Q* leak in an interview with The Verge. Some conspiracy theorists are putting forth the idea that Q* has figured out a way to break encryption, and OpenAI has tried to warn the NSA about it.

G42, an AI firm controlled by the United Arab Emirates (UAE), is facing scrutiny from US intelligence agencies over its connections with China. Despite a public image of rapid growth and partnerships with global companies like AstraZeneca and Silicon Valley firms, concerns have been raised that G42 is advancing Chinese AI interests in spite of US sanctions. The CIA and other US agencies have warned about G42's collaboration with major Chinese companies, particularly tech giant Huawei, which has been heavily sanctioned for years due to unscrupulous business practices.

These partnerships are viewed as potential threats to US security, with fears that G42 could serve as a conduit for advanced American technology to be transferred to Chinese companies or the government.

Highlighted were Q’s capabilities in content generation, taking actions, and understanding specific business needs by connecting to various apps and software like Salesforce, Jira, Zendesk, Gmail, and Amazon S3 storage. It can create service tickets, notify teams in Slack, and update dashboards in ServiceNow. Q also integrates with CodeWhisperer, Amazon's service for generating and interpreting app code. Also announced was an expansion to AWS’ partnership with Nvidia.

The Pentagon's AI initiatives are reshaping U.S. military strategy, with a focus on deploying thousands of lethal, AI-enabled autonomous vehicles by 2026. The ‘Replicator’ initiative, part of a broad effort to keep pace with China, represents a shift towards leveraging smaller, smarter, and more cost-effective platforms in warfare. AI technology is already being used in many US military applications, from piloting surveillance drones to predicting maintenance needs for Air Force planes - but lethal autonomy at scale will be a new frontier.

The ongoing war between Russia and the Ukraine has been the first real proving ground for AI, and there’s little doubt that the technology will shape the future of warfare.

Singapore's Changi Airport is testing an AI-driven system to speed up passenger security checks. The system uses AI and machine learning to interpret X-ray images, reducing processing time and potential human error. Initial results show the system performs as well as or better than human screeners in detecting prohibited items.

More in AI this week:

• Reshaping the tree: rebuilding organizations for AI

• Margaret Mitchell: ‘The people who are most likely to be harmed by AI don’t have a seat at the table for regulation’

• British AI chip darling Graphcore pulls out of China as Nvidia rival becomes latest casualty of US export curbs

• A Spanish agency became so sick of models and influencers that they created their own with AI - and she’s raking in up to $11,000 a month

• ‘Make it more’: generative AI memes explode onto the internet

• GPT-4 is able to buy stuff on Amazon, researchers say

• OpenAI’s Custom Chatbots Are Leaking Their Secrets

• What happens when your AI girlfriend dies?

• Pioneering AI technology diagnoses autism in children under two with 98.5% accuracy

• Sports Illustrated published articles by fake, AI-generated writers

Keep Your SSN Off The Dark Web

Data brokers play a shady game — harvesting your personal data online, and selling it to the highest bidder. No informed consent required.

In fact, every time you accept cookies or download an app from a bad actor, your data could be at risk. Opting out is possible of course — but made difficult by design.

It’d take the average person 304 hours to manually complete all the data removal requests needed to secure their privacy. That’s why the VPN specialists at Surfshark created Incogni: The data removal service that takes care of wiping your personal info from hundreds of databases, reducing the threat of robocalls, scams, and fraud.

Safeguarding your data shouldn’t take up your spare time. Leave it to Incogni — Our readers get 60% off the Incogni annual plan at this link with code PRIVACY.

See you next week!