Pharma AI Weekly #35: OpenAI’s o3-mini, Alibaba’s Qwen 2.5-Max, and New AI Copyright Protection Updates
Welcome to Pharma AI Weekly!
At ctcHealth, our team of experts understands that keeping up with AI advancements is crucial. That's why we're here to keep you informed and empowered!
Each week, our team carefully selects the most pertinent AI-related articles from the pharmaceutical industry, providing concise summaries and direct links to these valuable resources.
We invite you to Subscribe Here and discover how AI is transforming our industry.
Now, let’s explore three key AI developments impacting our industry this month:
Pharma AI Weekly Key Insights
1. OpenAI Hits Back at DeepSeek with o3-mini Reasoning Model
Published: January 31, 2025
OpenAI has introduced o3-mini, an AI model designed for reasoning and problem-solving. Unlike previous versions, this is the first OpenAI simulated reasoning model available for free to all users.
Key Highlights:
Improved Performance and Accuracy: OpenAI reports a 39% reduction in major errors compared to o1-mini, with testers preferring o3-mini’s responses 56% of the time due to improved accuracy.
Expanded Access & Features: o3-mini is replacing o1-mini for paid OpenAI subscribers, with higher daily message limits and the introduction of a new “Reason” mode available to free users.
New Search Capabilities and AI Limitations: The model includes a prototype search function for retrieving real-time answers with links to relevant web sources, but it still performs poorly in self-improvement and automating research tasks.
Why This Matters:
Better AI for Complex Reasoning Tasks: o3-mini’s enhanced reasoning allows users to fine-tune their balance between speed and accuracy, making it more effective for STEM applications, including coding and scientific research.
More Users Can Access Advanced AI for Free: Unlike previous OpenAI models, this is the first simulated reasoning AI made available to all users, increasing adoption potential.
AI Research Capabilities Remain Limited: Despite improvements, OpenAI acknowledges that o3-mini cannot automate machine learning research or perform self-improvement, keeping it a tool for assistance rather than independent discovery.
Competition Math (AIME 2024)
Mathematics: With low reasoning effort, OpenAI o3-mini achieves comparable performance with OpenAI o1-mini, while with medium effort, o3-mini achieves comparable performance with o1. Meanwhile, with high reasoning effort, o3-mini outperforms both OpenAI o1-mini and OpenAI o1, where the gray shaded regions show the performance of majority vote (consensus) with 64 samples.
PhD-level Science Questions (GPQA Diamond)
PhD-level science: On PhD-level biology, chemistry, and physics questions, with low reasoning effort, OpenAI o3-mini achieves performance above OpenAI o1-mini. With high effort, o3-mini achieves comparable performance with o1.
Competition Code (Codeforces)
Competition coding: On Codeforces competitive programming, OpenAI o3-mini achieves progressively higher Elo scores with increased reasoning effort, all outperforming o1-mini. With medium reasoning effort, it matches o1’s performance.
Read More: OpenAI hits back at DeepSeek with o3-mini reasoning model
Additional Resources:
2. Alibaba Releases AI Model It says Surpasses DeepSeek
Published: January 29, 2025
Alibaba has released Qwen 2.5-Max, claiming it outperforms DeepSeek-V3, GPT-4o, and Meta’s Llama-3.1-405B in various AI benchmarks. This move reflects China’s increasing AI competition, driven by DeepSeek’s disruptive low-cost models.
Key Highlights:
Competitive Performance in AI Benchmarks: Alibaba claims that Qwen 2.5-Max outperforms GPT-4o, DeepSeek-V3, and Llama-3.1-405B, suggesting improvements in complex reasoning, medical data processing, and regulatory compliance tasks.
DeepSeek’s Open-Source Model Disrupts AI Economics: DeepSeek’s low-cost AI models have forced Alibaba, ByteDance, and Baidu to drastically cut prices—up to 97% in some cases, making AI more affordable for pharma R&D.
New AI Developments Beyond OpenAI: ByteDance’s AI model claims superior performance in complex instruction-following benchmarks, highlighting increasingly viable alternatives to OpenAI for pharmaceutical applications.
Why This Matters:
Reduced AI Costs for Drug Development & Research: With AI prices dropping, pharmaceutical companies can now afford large-scale AI processing, improving drug discovery, patient data analysis, and clinical simulations.
More AI Model Choices for Regulatory Compliance: The emergence of Chinese AI providers with strong performance in reasoning and data interpretation presents new options for pharma companies needing AI-driven compliance solutions.
Growing Need for AI Strategy & Risk Assessment: Companies using AI models from Chinese providers must ensure regulatory compliance with HIPAA, GDPR, and FDA standards, requiring careful AI vendor evaluation.
Read More: Alibaba releases AI model it says surpasses DeepSeek
Additional Resources:
Alibaba Releases Its Own AI Model, Claiming it Rivals DeepSeek-V3
Alibaba Releases its AI Model Qwen 2.5 Claiming to Surpass DeepSeek
3. New Copyright Protection For AI, Plus New AI Models From Pica, Hailuo, And Bytedance
Published: January 30, 2025
The U.S. Copyright Office has issued new guidance stating that AI-generated content can only be copyrighted if it involves human creativity. This decision has direct implications for AI-generated research, regulatory submissions, and digital media.
Key Highlights:
AI-Generated Content Must Show Human Input to Be Copyrighted: The ruling clarifies that AI-generated works cannot receive copyright protection unless a human provides meaningful creative input.
Potential Impacts on AI-Generated Research and Documentation: Scientific papers, medical literature, and regulatory filings must have clear human contributions to retain intellectual property rights.
New AI Tools Are Expanding Content Creation Capabilities: Emerging AI tools like Pika 2.1 (high-resolution video generation), T2V-01-Director (camera movement AI), and INFP (AI-driven speech animation) highlight how AI is increasingly being used in digital content production.
Why This Matters:
AI-Assisted Research Needs Human Oversight for Legal Protection: If AI is used in scientific discovery, drug development, or documentation, companies must ensure human oversight to maintain intellectual property rights.
New Legal Considerations for AI-Generated Documents and Reports: Any AI-generated content submitted for regulatory or business purposes may require human review to ensure compliance with copyright and intellectual property laws.
AI’s Role in Content Creation Continues to Grow: As AI becomes more advanced, organizations must balance efficiency with legal compliance to maximize AI’s benefits while ensuring proper attribution and protection.
Read More: New Copyright Protection For AI, Plus New AI Models From Pica, Hailuo, And Bytedance
Additional Resources: Copyright Office Offers Assurances on AI Filmmaking Tools
Did you find this helpful?
If you enjoy the read, subscribe on LinkedIn to this weekly newsletter to stay updated on Pharma AI! Feel free to contact us at ctcHealth and explore how you can harness the power of AI in your sales operations.
Best regards,
The ctcHealth Team
About ctcHealth
Founded by Thomas Mrosk in 2024, ctcHealth specializes in AI Automation Consulting and AI Coaching services designed to transform the sales process in the healthcare industry.
Our diverse team combines pharmaceutical expertise and cutting-edge AI to create tailored solutions that meet today's market demands.