Overclocked
Posts
AI Could Be First to Detect and Contact Aliens

AI Could Be First to Detect and Contact Aliens

Zoe Kopidis
June 09, 2025

Welcome to this week's edition of Overclocked!

In this issue, we uncover how neural networks sift through cosmic noise to flag potential technosignatures (i.e. aliens), and we examine AI’s rapidly growing role in healthcare. Let’s dive in! ⬇️

In today’s newsletter ↓
👽 Will AI help us find aliens?
🧑‍⚕️ AI is transforming the face of healthcare
🧑‍⚖️ Open AI challenges court order to retain user data
🪫 Meta goes nuclear for the next 20 years
🔍 Weekly Challenge: AI Tool Selection

👽 Can AI Help Speed up the Search for Aliens?

AI is no longer science fiction in the search for extraterrestrial intelligence (SETI); it's a critical tool revolutionizing the field. By automating the analysis of immense cosmic datasets, AI is dramatically accelerating our hunt for alien technosignatures.

🤖 AI in Action

Astronomers are now deploying sophisticated AI to sift through data from powerful telescopes. At the forefront is Harvard's Galileo Project, which uses machine-learning pipelines for anomaly detection. Similarly, the SETI Institute and Breakthrough Listen utilize convolutional neural networks (CNNs), a type of AI originally developed for image recognition, to scan spectrograms for non-natural signals.

Presenting UAP Class 1 – Tetra.
This class is characterized by a tumbling or rotation-like motion, observed mid-air without any visible means of propulsion. Coordinated formations are also common, appearing consistently across both electro-optical (EO) and infrared (IR) sensors.
— Skywatcher (@SkywatcherHQ)
10:00 PM • May 28, 2025

These systems are trained on millions of labeled observations, from pulsars to human-made satellite interference. By learning to identify and filter out known signals, the AI can flag faint or unusual patterns that human analysts might miss, significantly reducing false positives and focusing research on the most promising candidates.

🔭 The Coming Data Tsunami

The need for AI will only grow as next-generation observatories come online. The SKA Observatory (SKAO), for instance, will soon generate exabytes of data daily—an amount impossible for humans to analyze manually.

Experience the first image from the SKA-Low telescope in a completely new way! ✨
🎶 For Global Accessibility Awareness Day, we've created a sonification of the dozens of radio galaxies spread across the image. Bright galaxies sound loud and high, dimmer galaxies soft and low.
— SKA Observatory (@SKAO)
7:10 AM • May 15, 2025

Only AI algorithms can process this deluge in real time, triaging potential signals and directing follow-up observations. This automated approach will be essential for correlating transient events across the sky and optimizing our search strategies, turning a needle-in-a-haystack problem into a manageable data-sifting task.

⚖️ Ethical and Legal Frontiers

The rapid advancement of AI also introduces new challenges. While not directly related to telescope logs, the recent lawsuit filed by Reddit against AI firm Anthropic over alleged unauthorized data scraping to train large language models highlights a growing concern.

As astronomical data becomes increasingly valuable, clear regulations will be needed to govern its use in training AI, ensuring that the search remains both ethical and collaborative. AI is a powerful partner, but its deployment requires careful consideration. If we do make contact, it may very well be an AI that first alerts us.

🩺 AI Is Changing Healthcare Forever

Hospitals are deploying deep-learning tools that read medical images alongside clinicians. According to U.S. News & World Report, AI systems have matched dermatologists in flagging suspicious moles, and early studies show similar gains in mammography and retinal scans for diabetes care. When algorithms prescreen images, radiologists can focus on borderline cases—speeding up diagnoses without replacing human oversight.

🗂️ Cutting the Paperwork Burden

Generative AI isn’t just for imaging. Primary-care networks now test speech-to-text “ambient scribes” that draft visit notes and suggest billing codes in real time. Lexology reports that physicians using these assistants reclaim up to two hours per day once spent on electronic health records, leading to higher patient-throughput and lower burnout. Automatic summarization also helps specialists sift through decades of fragmented charts before a single consult.

Doctors, clinicians, and medical experts - time is your most valuable asset.
With AI, you can save hours every day.
Here’s how:
— Chidanand Tripathi (@thetripathi58)
10:40 AM • Jun 6, 2025

🌐 Toward Precision & Prevention

McKinsey & Company estimates that applying generative AI across drug discovery, care delivery, and back-office functions could unlock $1 trillion of annual value worldwide. Early pilots pair large-language models with hospital data lakes to surface at-risk patients—such as predicting sepsis hours before symptoms escalate—while pharma teams use the same tech to generate hypothesis-ready target lists and draft regulatory documents.

⚖️ Guardrails Still Forming

AI medical devices approved by the U.S. FDA remain “locked”: any major model update triggers a new review. Liability questions linger—who is responsible if an AI-suggested diagnosis proves harmful? Regulatory bodies are drafting guidelines that require human sign-off and transparent audit logs, ensuring clinicians—not algorithms—retain the final word.

Bottom line: AI is becoming a clinical co-pilot—triaging scans, drafting notes, and hinting at therapy paths—yet trust hinges on rigorous validation and clear accountability. Hospitals that blend algorithmic speed with human judgment are already seeing gains in efficiency and patient outcomes.

The Weekly Scoop 🍦

🧠 OpenAI challenges a court mandate requiring it retain all user data

⚖️ Reddit sues Anthropic for freeloading on its user data

⚡ Meta locked in on nuclear for the next 20 years

📊 Google’s AI Mode unlocks interactive financial charts on the go

🧬 AI makes radiologists 40-50% more efficient

🎯 Weekly Challenge: Pick Your Power Tool

Challenge: Think you know your AI tools? Let’s put that to the test. Below are 5 real-world tasks—and your mission is to match the job to the tool you think would crush it best.

🧠 Task 1: Scientific Sleuthing
You need to find a 2024 study on using AI to predict volcanic eruptions.
Which AI gives you a structured summary, key metrics, and citations in under 2 minutes?