🤖 TikTok Fires Intern Who “Maliciously Interfered” With AI Research

Subscribe | AI Toolkit | Meet The Team

Welcome back AI enthusiasts!

In today’s AI Report:

• 👨‍🚒TikTok Fires Intern Who “Maliciously Interfered” With AI Research

• 🦺Anthropic’s Four New Sabotage Evaluations for Advanced AI Models

• 📊“SymGen” Verifies AI Model Responses

• 🛠Trending Tools

• 💰Funding Frontlines

• 💼Who’s Hiring?

Read Time: 3 minutes

🗞RECENT NEWS

BYTEDANCE

👨‍🚒TikTok Fires Intern Who “Maliciously Interfered” With AI Research

TikTok’s parent company, ByteDance, confirmed it had to fire an intern who “maliciously interfered” with TikTok’s AI research.

Key Details:

• A Commercial Technology intern “committed serious disciplinary violations,” said ByteDance. “The intern maliciously interfered with AI model training.”

• The intern’s actions affected ByteDance’s AI Training Program, where employees program an AI model by “training” it on vast amounts of data to recognize patterns, understand context, and make decisions.

• The intern allegedly implanted an Unsafe Pickle (i.e., malicious code injection) that actively “untrained” AI models.

• The intern would participate in meetings where they tried to solve the issues, allowing him to adapt his strategies and avoid detection.

• ByteDance denies the claims that the intern impacted 8,000 Graphics Processing Units (GPUs) or cost TikTok “tens of millions of dollars.”

• For context, Nvidia H100 Tensor Core GPUs are estimated to cost around $25,000 per GPU.

🚨Read the latest updates on this situation here.

🩺 PULSE CHECK

ANTHROPIC

🦺Anthropic’s Four New Sabotage Evaluations for Advanced AI Models

Anthropic, a research company building reliable, interpretable, and steerable AI systems, released four new Sabotage Evaluations for advanced AI models.

Key Details:

• In theory, advanced AI models could “subvert human oversight and decision-making in important contexts.”

• In other words, as AI models become more sophisticated, they could make decisions or take actions humans don’t want.

• For example, advanced AI models could “covertly sabotage efforts to evaluate their own dangerous capabilities, to monitor their behavior, or to make decisions about their deployment.”

• The four new Sabotage Evaluations include Code, Sandbagging, Human Decisions, and Undermining Oversight.

• They examine an advanced AI model’s capabilities to steer humans toward bad decisions without appearing suspicious, insert bugs into codebases, and systematically undermine monitoring procedures.

Why It’s Important:

• Anthropic is open-sourcing the four new Sabotage Evaluations because they “hope other AI researchers will use, critique, and improve upon” their work.

• Uncovering potential vulnerabilities that might not be apparent in standard AI model testing helps to identify areas where human oversight is necessary.

AI RESEARCH

📊“SymGen” Verifies AI Model Responses

Despite their impressive capabilities, AI models are far from perfect. They sometimes “hallucinate” by confidently generating inaccurate or misleading information.

To prevent this, an AI model’s responses are verified by human fact-checkers. However, this error-prone process requires them to read through extensive reports filled with citations manually.

To solve this issue, MIT researchers developed “SymGen,” a user-friendly framework that enables anyone to verify an AI model’s responses in minutes. “SymGen” generates responses with citations directly referring to the source.

“SymGen” streamlines the verification process for human fact-checkers by around 20%. By making the validation of an AI model’s responses easier and faster, “SymGen” could prevent errors with AI applications deployed in safety-critical healthcare situations.

“It can give people higher confidence in an AI model’s responses because they can easily take a closer look to ensure the information is verified,” said Shannon Zejiang Shen, co-lead author of the paper on “SymGen.”

🛠TRENDING TOOLS

🗣Convo is the most powerful qualitative research platform.

📈TradingLiteracy lets you chat with your Trade History files.

⏳Feta plans and drives meetings that’re worth everyone’s time.

🚜ToolBuilder creates AI tools effortlessly with a single prompt.

👟Kick is an AI-assisted account software that does the work for you.

🔮Browse our always Up-To-Date AI Tools Database.

💰FUNDING FRONTLINES

• GenAI investments topped $3.9B in Q3 FY2024.

• Eruditus bags a $150M Series F at a $3.1B valuation to improve how students learn with AI.

• Cyera acquires Trail Security for $162M to reimagine AI-powered data security solutions.

💼WHO’S HIRING?

• Sigma Computing (San Francisco, CA): Software Engineering Intern, Summer 2025

• GM Financial (Arlington, TX): Software Development Engineer Intern, Summer 2025

• The New York Times {NYT} (New York, NY): Backend Engineering Intern, Summer 2025

• Adobe (San Jose, CA): Research Scientist/Engineer, New College Grad 2025

• Apple (Austin, TX): Machine Learning {ML} Engineer, Applied Data Science Program, Early Career

🤖PROMPT OF THE DAY

MARKETING

🛒Outline Marketing Strategies

Outline an effective marketing plan for [Small Business] in [Industry] to launch [Product/Service] aimed at [Target Audience].

Make sure to include the “Four Ps”: a product, price, place, and promotion.

Small Business = [Insert Here]

Industry = [Insert Here]

Product/Service = [Insert Here]

Target Audience = [Insert Here]

📒FINAL NOTE

FEEDBACK

❤️TAIP Review of The Day

REFER & EARN

🎉Your Friends Learn, You Earn!

You currently have 0 referrals, only 1 away from receiving ⚙️Ultimate Prompt Engineering Guide.

Refer 3 friends to learn how to 👷‍♀️Build Custom Versions of OpenAI’s ChatGPT.

Copy and paste this link to friends: https://theaipulse.beehiiv.com/subscribe?ref=PLACEHOLDER