Good morning, it’s Wednesday. Earth Day may be behind us, but the connection between AI and nature is just getting started. Meanwhile, Hollywood embraces AI, and AI “employees” may be closer than you think.

In today’s Forward Future Original, we explore how OpenAI’s o3 model uses tools autonomously to tackle complex, multimodal tasks.

Read on.

P.S. Hunting for your next AI gig—or just curious what’s out there? Don’t miss the new job listings at the end of today’s edition.

🗞️ YOUR DAILY ROLLUP

^{Top Stories of the Day}

🏆 Oscars Approve AI Use in Film Awards
The Academy now permits films using AI to compete for Oscars, stating AI won’t impact a film’s eligibility either way. This follows the use of generative AI to enhance performances in recent Oscar-winning films. While the Academy emphasizes human input remains key, concerns persist among actors and writers over job security and ethical misuse. Industry safeguards are in place, but skepticism around AI’s creative limits remains strong.

👥 Anthropic: AI “Employees” Could Arrive by 2026
Anthropic predicts fully autonomous AI employees could hit corporate networks within a year, raising urgent cybersecurity questions. These AI agents would hold roles, retain memories, and access internal systems with their own credentials. CISO Jason Clinton warns that without rethinking security models, companies risk major breaches. Tools to monitor and manage these non-human identities are already emerging—but clarity on accountability remains elusive.

💰 Google Pays Samsung Big for Gemini Preinstalls
Google has been paying Samsung hefty monthly sums since January to preinstall its Gemini AI app on devices, a move now under scrutiny in a U.S. antitrust case. Testimony reveals the deal includes fixed payments and ad revenue sharing, despite similar practices being ruled illegal in the past. Previously, Google paid Samsung $8B over three years for default placements of its apps.

🚗 Rivian Taps Cohere CEO for Board, Doubling Down on AI
EV maker Rivian has appointed Cohere CEO Aidan Gomez—co-author of the foundational AI paper “Attention Is All You Need”—to its board. The move comes as Rivian doubles down on AI, including a $5.8B software venture with VW and development of an in-car assistant. It’s a clear signal: Rivian sees generative AI as central to its future.

Enjoying our newsletter? Forward it to a colleague—
it’s one of the best ways to support us.

🌿 CONSERVATION

AI’s Role in Wildlife Conservation Sparks Promise—and Controversy

The Recap: AI is accelerating biodiversity research by processing massive volumes of wildlife data, from counting pelicans in Senegal to uncovering unknown species in Panama. While scientists hail its efficiency and new capabilities, critics warn of ethical blind spots, environmental costs, and potential biases baked into its use.

Highlights:

• AI systems like object detection models are used to count species in aerial footage and camera trap images, significantly reducing human labor and time.

• Alexandre Delplanque’s “HerdNet” model helped count elephants and antelopes on the savannah and required only 12 hours of local computer processing.

• Researchers in Panama used AI to identify over 300 previously undocumented species in just a week’s worth of camera trap footage.

• Critics like Hamish van der Ven and Shaolei Ren highlight AI’s environmental toll—ranging from energy usage and water consumption to local air pollution near data centers.

• Projects like the Earth Species Project and Project CETI aim to decode animal communication using generative AI and LLMs, raising ethical and scientific concerns.

Forward Future Takeaways:
The use of AI in wildlife conservation reflects a broader dilemma: how to deploy powerful tools ethically and sustainably. While AI enables unprecedented insight into biodiversity, its environmental footprint and social implications demand scrutiny. As these tools evolve, the real test will be whether AI enhances conservation without undermining the ecosystems it aims to protect — and whether we ask the right ecological questions, not just the convenient ones. → Read the full article here.

👾 FORWARD FUTURE ORIGINAL

Why Tool Use Marks a New Era for AI

Imagine an artificial intelligence that not only answers questions, but also independently researches, analyzes data, interprets images and makes informed decisions - all in a fluid process. With the introduction of OpenAI's o3 model in April 2025, this has become a reality. This model marks a turning point in AI development by being able to autonomously use a variety of tools to accomplish complex tasks for the first time. → Continue reading here.

📏 BENCHMARKS

Experts Warn That Crowdsourced AI Benchmarks May Mislead More Than They Measure

The Recap: AI researchers and ethicists are raising concerns about the validity and ethics of crowdsourced benchmarking platforms like Chatbot Arena. Critics argue that these benchmarks lack scientific rigor, are prone to misuse by labs, and risk undervaluing evaluators’ labor. The article, reported by Kyle Wiggers, includes perspectives from academics, AI founders, and platform creators who call for deeper scrutiny and structural reform.

Highlights:

• AI labs including OpenAI, Google, and Meta are increasingly using crowdsourced platforms like Chatbot Arena to evaluate model performance.

• Linguist Emily Bender criticized Chatbot Arena for lacking “construct validity,” stating it fails to prove that user preferences reflect meaningful or measurable model quality.

• Asmelash Teka Hadgu accused Meta of gaming the system by fine-tuning a high-scoring Llama 4 variant for benchmarks, then withholding it from release.

• Hadgu advocates for dynamic, domain-specific benchmarks led by independent institutions and professionals in applied fields like healthcare and education.

• LMArena co-founder Wei-Lin Chiang defended Chatbot Arena’s intent and transparency, citing recent policy updates to prevent benchmark discrepancies and maintain trust.

Forward Future Takeaways:
This debate strikes at the heart of how we assess progress in AI: if benchmarks become marketing tools rather than scientific measures, the field risks distorting both public perception and internal innovation. Crowdsourced evaluation can offer valuable insights, but without rigorous standards and fair labor practices, it may fall short of its promise. As AI systems continue to shape real-world decisions, how we measure their capabilities matters more than ever. → Read the full article here.

🛰️ NEWS

What Else is Happening

💸 Manychat Scores $140M Boost: The chat platform plans to scale AI features and global reach, already serving 1.5M users across 170 countries.

🤖 DIA Launches Open-Source Speech Model: New TTS challenger takes on ElevenLabs and OpenAI with lifelike voices, transparency, and open training data.

🏗️ NVIDIA CEO Pushes Japan on AI: Jensen Huang urges LDP to invest in domestic AI infrastructure, calling it key to Japan’s robotics and tech future.

📟 Adaptive Computer Debuts ‘Vibe’ Coding: Backed by $7M, the startup lets non-coders build full apps via text prompts—no API keys or dev skills needed.

🤓 AI Tackles Global Nearsightedness Surge: New models help detect and predict myopia early, but challenges in data quality and clinical trust remain.

 🔬 RESEARCH

AI Gives Old Solar Data a New Glow

A new AI-powered method is helping solar scientists unify decades of fragmented sun data, translating observations from outdated instruments into a format compatible with modern tools. Developed by researchers in Austria and Russia, the technique uses dual neural networks to mimic and reverse the degradation caused by older telescopes—effectively upgrading historical solar images without losing key physical details. → Read the full story here.

💼 JOB BOARD

Now Hiring: Anthropic, ElevenLabs, & More

🧰 TOOLBOX

Smarter Code Reviews, Simpler Compliance, and No-Code AI Workflows

⚙️ Trag AI: Automate code reviews with AI—catch bugs early, enforce standards, and get real-time feedback across any language or repo.

🍪 AWS Cookie Preferences: Simplify cookie consent on AWS-hosted sites—auto-generate banners, manage user preferences, and stay compliant with ease.

👨‍💻 Agent Network: Build AI agent workflows using natural language—customize tasks, chain actions, and run logic without writing complex code.

👉️ Find trending AI tools: Browse the Forward Future AI Tool Library

That’s a Wrap!

🛰️ Want more Forward Future? Follow us on X for quick updates, subscribe to our YouTube for deep dives, or add us to your RSS Feed for seamless reading.

Thanks for reading today’s newsletter—see you next time!

The Forward Future Team

🧑‍🚀 🧑‍🚀 🧑‍🚀 🧑‍🚀