AI Secrets is your new weekly update on how AI is changing our everyday lives. Our experts will keep it clear and simple, so you can stay ahead of the game. This week we’re looking at Next-gen AIs (and the end of humanity).Share this with anyone you want to keep up to date.
Maybe we just need a break? 🚩🚩
It’s been another wild, hype-filled week for the AI world. It began with Microsoft claiming GPT-4 was showing “sparks” of Artificial General Intelligence, or AGI; the term used for AIs that can perform any mental task as well as a human. It ended with a raft of big names in the tech industry - including OpenAI co-founder Elon Musk and Apple co-founder Steve Wozniak - signing an open letter from the Future Of Life Institute warning that, in the long term, AIs pose an existential risk to humanity. The letter calls for a six-month pause on the development of new AI systems “more powerful than GPT-4”, allowing regulators time to implement the necessary safeguards and save the human race from a robopocalypse.
The headlines write themselves, but many in the AI industry are skeptical of both of these claims, and that includes campaigners for more transparency and safeguarding. Embarrassingly for the Future of Life Institute, one of the loudest critics of their open letter is also one of the main academics cited in their work, Emily M. Bender. She took to Twitter to lambast the letter as “dreck” and “#AIhype”, and many critics have suggested that the letter’s main goal was to distract attention from the problems being posed by GPT-4 in 2023.
Meanwhile, Microsoft’s bold claim that they see “sparks of AGI” in GPT-4 has been met with similar accusations of unwarranted hype. These claims did not come in a press release but in a 155-page research paper published on the popular pre-print site arXiv, where academic papers can be posted before they go through the rigors of peer review. Emily Bender has described arXiv as a “cesspool in AI - a reservoir for hype infections”, and many have suggested that Microsoft’s submission is mainly intended to promote their new (GPT-4 equipped) Bing Search.
Behind all the controversy, the cause for concern is real. While debate rages over how far away AGI is - or even whether it is achievable at all - most insiders agree that unchecked AI development could lead to catastrophic outcomes in the long term. In 2021 researchers estimated the chance of an AI-driven “existential catastrophe” occurring by 2070 was 5%; last year’s spurt of new, far more powerful AIs caused them to revise that figure to 10%, and the pace of development continues to exceed expectations in 2023. Forewarned is forearmed, of course and, as such you should definitely check out our detailed explainer on AGI!
Special Limited Edition!
Puzzle lovers will adore this unique custom cube. Each of its six sides features a different design. With 43 quintillion unique positions it’s more complex than any spy story! Get ready to have fun and be challenged.
Not content with text-to-image AI apps like Midjourney, text-to-video AI apps have started appearing everywhere, and that can only mean one thing: unsettling deepfakes of Will Smith eating pasta. Do you dare watch?
Google Bard’s awkward launch lurches from crisis to crisis. Journalists are reporting that a prominent Google staffer resigned because Bard was being trained using OpenAI’s data - a charge Google flatly denies.
OpenAI’s ChatGPT’s outage last week was more serious than initially thought; the bug exposed the personal data of 1.2% of users, including address and credit card information.
AI continues to shake up the infosec world as Microsoft has launched Microsoft Copilot, an AI coding assistant specifically tailored to provide rapid responses to combat cybercrime and network vulnerabilities.
This week’s biggest viral AI star was none other than Pope Francis, after Midjourney deepfakes of him wearing a swagged-out ankle-length Balenciaga puffer jacket fooled the world.
Video meeting app Zoom is among the latest firms to bolt AI tech onto their product, which can now summarize meetings for you when you arrive late, and even ask questions on your behalf.
